Release v0.9.4

This is an edge case of how the lexer changes behavior when
reading string literals versus reading unquoted source code.
Three double quotes would begin a multi-line string literal,
but just two are still ambiguous, so it has to `peek()` to
see whether it's inside a multi-line string literal or
outside an empty regular string.

.clang-tidy

@@ -0,0 +1,6 @@
+---
+Checks: ''
+WarningsAsErrors: ''
+HeaderFilterRegex: ''
+FormatStyle: none
+SystemHeaders: false

.github/scripts/build_libpng.sh

@@ -1,13 +1,13 @@
 #!/bin/bash
 set -euo pipefail
 
-pngver=1.6.45
+pngver=1.6.50
 
 ## Grab sources and check them
 
 curl -LOJ "http://prdownloads.sourceforge.net/libpng/libpng-$pngver.tar.xz?download"
 # Brew doesn't provide any sha256sum, so we're making do with `sha2` instead.
-if [ "$(sha2 -q -256 libpng-$pngver.tar.xz)" != 926485350139ffb51ef69760db35f78846c805fef3d59bfdcb2fba704663f370 ]; then
+if [ "$(sha2 -q -256 libpng-$pngver.tar.xz)" != 4df396518620a7aa3651443e87d1b2862e4e88cad135a8b93423e01706232307 ]; then
 	sha2 -256 libpng-$pngver.tar.xz
 	echo Checksum mismatch! Aborting. >&2
 	exit 1

.github/scripts/get_win_deps.ps1

@@ -16,8 +16,8 @@ function getlibrary ([string] $URI, [string] $filename, [string] $hash, [string]
 }
 
 getlibrary 'https://www.zlib.net/zlib131.zip' 'zlib.zip' '72af66d44fcc14c22013b46b814d5d2514673dda3d115e64b690c1ad636e7b17' .
-getlibrary 'https://github.com/glennrp/libpng/archive/refs/tags/v1.6.45.zip' 'libpng.zip'  '1b3d94b2f1d137db1bf1842cb9f03df179772a517f7b86e26351742190632785' .
+getlibrary 'https://github.com/pnggroup/libpng/archive/refs/tags/v1.6.50.zip' 'libpng.zip' 'f6bb2544d2cf5465af3a695dee0b7eacff82f11a50aa4672ef0e19df6e16d455' .
 getlibrary 'https://github.com/lexxmark/winflexbison/releases/download/v2.5.25/win_flex_bison-2.5.25.zip' 'winflexbison.zip' '8d324b62be33604b2c45ad1dd34ab93d722534448f55a16ca7292de32b6ac135' install_dir
 
 Move-Item zlib-1.3.1 zlib
-Move-Item libpng-1.6.45 libpng
+Move-Item libpng-1.6.50 libpng

.github/scripts/mingw-w64-libpng-dev.sh

@@ -1,22 +1,17 @@
 #!/bin/bash
 set -euo pipefail
 
-pngver=1.6.45
+pngver=1.6.50
 arch="$1"
 
 ## Grab sources and check them
 
 wget http://downloads.sourceforge.net/project/libpng/libpng16/$pngver/libpng-$pngver.tar.xz
-wget http://downloads.sourceforge.net/project/apng/libpng/libpng16/libpng-$pngver-apng.patch.gz
-sha256sum -c .github/scripts/mingw-w64-libpng-dev.sha256sums
+echo 4df396518620a7aa3651443e87d1b2862e4e88cad135a8b93423e01706232307 libpng-$pngver.tar.xz | sha256sum -c -
 
 ## Extract sources and patch them
 
 tar -xf libpng-$pngver.tar.xz
-gunzip libpng-$pngver-apng.patch.gz
-
-# Patch in apng support
-env -C libpng-$pngver patch -p0 ../libpng-$pngver-apng.patch
 
 ## Start building!
 

.github/scripts/mingw-w64-libpng-dev.sha256sums

@@ -1,2 +0,0 @@
-f7caa3b55f003ce23d6a087b1c2a643262647bfd1a1b31afdc9b18eabf1bbc7e  libpng-1.6.45-apng.patch.gz
-926485350139ffb51ef69760db35f78846c805fef3d59bfdcb2fba704663f370  libpng-1.6.45.tar.xz

.github/workflows/build-container.yml

@@ -37,6 +37,7 @@ jobs:
           TAG_NAME=${GITHUB_REF#refs/tags/}
           sed -i "2i LABEL org.opencontainers.image.description=\"RGBDS container image for the release version $TAG_NAME\"" Dockerfile
           docker build . --tag ghcr.io/gbdev/rgbds:$TAG_NAME
+          docker tag ghcr.io/gbdev/rgbds:$TAG_NAME ghcr.io/gbdev/rgbds:latest
           docker push ghcr.io/gbdev/rgbds:$TAG_NAME
           docker push ghcr.io/gbdev/rgbds:latest
 

.github/workflows/checkdiff.yml

@@ -1,4 +1,4 @@
-name: Code coverage checking
+name: Diff completeness check
 on: pull_request
 
 jobs:

.github/workflows/checkformat.yml

@@ -9,4 +9,4 @@ jobs:
         uses: actions/checkout@v4
       - name: Check format
         run: |
-          ./contrib/checkformat.bash
+          contrib/checkformat.bash

.github/workflows/coverage.yml

@@ -0,0 +1,34 @@
+name: Code coverage report
+on:
+  - push
+  - pull_request
+
+jobs:
+  coverage:
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Checkout repo
+        uses: actions/checkout@v4
+      - name: Install deps
+        shell: bash
+        run: |
+          ./.github/scripts/install_deps.sh ubuntu
+      - name: Install LCOV
+        run: |
+          sudo apt-get install lcov
+      - name: Install test dependency dependencies
+        shell: bash
+        run: |
+          test/fetch-test-deps.sh --get-deps ubuntu
+      - name: Generate coverage report
+        run: |
+          contrib/coverage.bash ubuntu-ci
+      - name: Upload coverage report
+        uses: actions/upload-artifact@v4
+        with:
+          name: coverage-report
+          # Workaround for keeping the top-level coverage/ directory
+          # https://github.com/actions/upload-artifact/issues/174
+          path: |
+            coverage
+            dummy-file-to-keep-directory-structure.txt

.github/workflows/testing.yml

@@ -69,7 +69,7 @@ jobs:
       - name: Run tests
         shell: bash
         run: |
-          CXX=${{ matrix.cxx }} test/run-tests.sh
+          CXX=${{ matrix.cxx }} test/run-tests.sh --os ${{ matrix.os }}
 
   macos-static:
     runs-on: macos-14
@@ -121,13 +121,13 @@ jobs:
       - name: Run tests
         shell: bash
         run: |
-          test/run-tests.sh
+          test/run-tests.sh --os macos
 
   windows:
     strategy:
       matrix:
         bits: [32, 64]
-        os: [windows-2019, windows-2022]
+        os: [windows-2022, windows-2025]
         include:
           - bits: 32
             arch: x86
@@ -149,9 +149,10 @@ jobs:
           path: |
             zbuild
             pngbuild
-          key: ${{ matrix.arch }}-${{ hashFiles('zlib/**', 'libpng/**') }}
+          key: ${{ matrix.os }}-${{ matrix.arch }}-${{ hashFiles('zlib/**', 'libpng/**') }}
       - name: Build zlib
         if: steps.cache.outputs.cache-hit != 'true'
+        shell: bash
         run: | # BUILD_SHARED_LIBS causes the output DLL to be correctly called `zlib1.dll`
           cmake -S zlib -B zbuild -A ${{ matrix.platform }} -Wno-dev -DCMAKE_INSTALL_PREFIX=install_dir -DBUILD_SHARED_LIBS=ON
           cmake --build zbuild --config Release -j
@@ -212,7 +213,7 @@ jobs:
         run: |
           cp bins/* .
           cp bins/*.dll test/gfx
-          test/run-tests.sh
+          test/run-tests.sh --os ${{ matrix.os }}
 
   windows-mingw-build:
     strategy:
@@ -270,7 +271,7 @@ jobs:
     needs: windows-mingw-build
     strategy:
       matrix:
-        os: [windows-2019, windows-2022]
+        os: [windows-2022, windows-2025]
         bits: [32, 64]
       fail-fast: false
     runs-on: ${{ matrix.os }}
@@ -319,7 +320,7 @@ jobs:
       - name: Run tests
         shell: bash
         run: |
-          test/run-tests.sh
+          test/run-tests.sh --os ${{ matrix.os }}
 
   cygwin:
     strategy:
@@ -331,7 +332,7 @@ jobs:
           - bits: 64
             arch: x86_64
       fail-fast: false
-    runs-on: windows-2019
+    runs-on: windows-2022
     timeout-minutes: 30
     steps:
       - name: Checkout repo
@@ -382,7 +383,7 @@ jobs:
               git \
               png
           run: | # FreeBSD `c++` compiler does not support `make develop` sanitizers ASan or UBSan
-            cmake -S . -B build -DCMAKE_BUILD_TYPE=Debug -DCMAKE_CXX_COMPILER=c++ -DUSE_EXTERNAL_TESTS=OFF
+            cmake -S . -B build -DCMAKE_BUILD_TYPE=Debug -DCMAKE_CXX_COMPILER=c++ -DUSE_EXTERNAL_TESTS=OFF -DOS=bsd
             cmake --build build -j4 --verbose
             cmake --install build --verbose
             cmake --build build --target test

CONTRIBUTING.md

@@ -97,7 +97,7 @@ All tests begin by assembling the `.asm` file into an object file, which will be
 
 These simply check that RGBLINK's output matches some expected output.
 
-A `.out` file **must** exist, and RGBLINK's output must match that file's contents.
+A `.out` file **must** exist, and RGBLINK's total output must match that file's contents.
 
 Additionally, if a `.out.bin` file exists, the `.gb` file generated by RGBLINK must match it.
 
@@ -106,14 +106,14 @@ Additionally, if a `.out.bin` file exists, the `.gb` file generated by RGBLINK m
 These allow applying various linker scripts to the same object file.
 If one or more `.link` files exist, whose names start the same as the `.asm` file, then each of those files correspond to one test.
 
-Each `.link` linker script **must** be accompanied by a `.out` file, and RGBLINK's output must match that file's contents when passed the corresponding linker script.
+Each `.link` linker script **must** be accompanied by a `.out` file, and RGBLINK's total output must match that file's contents when passed the corresponding linker script.
 
 #### Variant tests
 
 These allow testing RGBLINK's `-d`, `-t`, and `-w` flags.
 If one or more <code>-<var>&lt;flag&gt;</var>.out</code> or <code>-no-<var>&lt;flag&gt;</var>.out</code> files exist, then each of them corresponds to one test.
 
-The object file will be linked with and without said flag, respectively; and in each case, RGBLINK's output must match the `.out` file's contents.
+The object file will be linked with and without said flag, respectively; and in each case, RGBLINK's total output must match the `.out` file's contents.
 
 ### RGBFIX
 
@@ -123,8 +123,11 @@ Each one is a text file whose first line contains flags to pass to RGBFIX.
 
 RGBFIX will be invoked on the `.bin` file if it exists, or else on default-input.bin.
 
+If no `.out` file exist, RGBFIX is not expected to output anything.
+If one *does* exist, RGBFIX's output **must** match the `.out` file's contents.
+
 If no `.err` file exists, RGBFIX is simply expected to be able to process the file normally.
-If one *does* exist, RGBFIX's return status is ignored, but its output **must** match the `.err` file's contents.
+If one *does* exist, RGBFIX's return status is ignored, but its error output **must** match the `.err` file's contents.
 
 Additionally, if a `.gb` file exists, the output of RGBFIX must match the `.gb`.
 

Dockerfile

@@ -1,6 +1,6 @@
 FROM debian:12-slim
 LABEL org.opencontainers.image.source=https://github.com/gbdev/rgbds
-ARG version=0.9.3
+ARG version=0.9.4
 WORKDIR /rgbds
 
 COPY . .

LICENSE

@@ -1,6 +1,6 @@
 The MIT License
 
-Copyright (c) 1997-2024, Carsten Sørensen and RGBDS contributors.
+Copyright (c) 1996-2025, Carsten Sørensen and RGBDS contributors.
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to

Makefile

@@ -3,7 +3,7 @@
 .SUFFIXES:
 .SUFFIXES: .cpp .y .o
 
-.PHONY: all clean install checkdiff develop debug profile coverage iwyu mingw32 mingw64 wine-shim dist
+.PHONY: all clean install checkdiff develop debug profile coverage tidy iwyu mingw32 mingw64 wine-shim dist
 
 # User-defined variables
 
@@ -49,7 +49,14 @@ BASE_REF := origin/master
 
 all: rgbasm rgblink rgbfix rgbgfx
 
+common_obj := \
+	src/extern/getopt.o \
+	src/diagnostics.o \
+	src/usage.o
+
 rgbasm_obj := \
+	${common_obj} \
+	src/asm/actions.o \
 	src/asm/charmap.o \
 	src/asm/fixpoint.o \
 	src/asm/format.o \
@@ -64,9 +71,7 @@ rgbasm_obj := \
 	src/asm/section.o \
 	src/asm/symbol.o \
 	src/asm/warning.o \
-	src/extern/getopt.o \
 	src/extern/utf8decoder.o \
-	src/error.o \
 	src/linkdefs.o \
 	src/opmath.o \
 	src/util.o
@@ -74,7 +79,10 @@ rgbasm_obj := \
 src/asm/lexer.o src/asm/main.o: src/asm/parser.hpp
 
 rgblink_obj := \
+	${common_obj} \
 	src/link/assign.o \
+	src/link/lexer.o \
+	src/link/layout.o \
 	src/link/main.o \
 	src/link/object.o \
 	src/link/output.o \
@@ -83,34 +91,36 @@ rgblink_obj := \
 	src/link/sdas_obj.o \
 	src/link/section.o \
 	src/link/symbol.o \
-	src/extern/getopt.o \
+	src/link/warning.o \
 	src/extern/utf8decoder.o \
-	src/error.o \
 	src/linkdefs.o \
 	src/opmath.o \
 	src/util.o
 
-src/link/main.o: src/link/script.hpp
+src/link/lexer.o src/link/main.o: src/link/script.hpp
 
 rgbfix_obj := \
+	${common_obj} \
 	src/fix/main.o \
-	src/extern/getopt.o \
-	src/error.o
+	src/fix/mbc.o \
+	src/fix/warning.o
 
 rgbgfx_obj := \
+	${common_obj} \
+	src/gfx/color_set.o \
 	src/gfx/main.o \
 	src/gfx/pal_packing.o \
 	src/gfx/pal_sorting.o \
 	src/gfx/pal_spec.o \
+	src/gfx/png.o \
 	src/gfx/process.o \
-	src/gfx/proto_palette.o \
 	src/gfx/reverse.o \
 	src/gfx/rgba.o \
-	src/extern/getopt.o \
-	src/error.o
+	src/gfx/warning.o \
+	src/util.o
 
 rgbasm: ${rgbasm_obj}
-	$Q${CXX} ${REALLDFLAGS} -o $@ ${rgbasm_obj} ${REALCXXFLAGS} src/version.cpp -lm
+	$Q${CXX} ${REALLDFLAGS} -o $@ ${rgbasm_obj} ${REALCXXFLAGS} src/version.cpp
 
 rgblink: ${rgblink_obj}
 	$Q${CXX} ${REALLDFLAGS} -o $@ ${rgblink_obj} ${REALCXXFLAGS} src/version.cpp
@@ -142,6 +152,8 @@ src/link/script.hpp: src/link/script.cpp
 	$Qtouch $@
 
 # Only RGBGFX uses libpng (POSIX make doesn't support pattern rules to cover all these)
+src/gfx/color_set.o: src/gfx/color_set.cpp
+	$Q${CXX} ${REALCXXFLAGS} ${PNGCFLAGS} -c -o $@ $<
 src/gfx/main.o: src/gfx/main.cpp
 	$Q${CXX} ${REALCXXFLAGS} ${PNGCFLAGS} -c -o $@ $<
 src/gfx/pal_packing.o: src/gfx/pal_packing.cpp
@@ -150,9 +162,9 @@ src/gfx/pal_sorting.o: src/gfx/pal_sorting.cpp
 	$Q${CXX} ${REALCXXFLAGS} ${PNGCFLAGS} -c -o $@ $<
 src/gfx/pal_spec.o: src/gfx/pal_spec.cpp
 	$Q${CXX} ${REALCXXFLAGS} ${PNGCFLAGS} -c -o $@ $<
-src/gfx/process.o: src/gfx/process.cpp
+src/gfx/png.o: src/gfx/png.cpp
 	$Q${CXX} ${REALCXXFLAGS} ${PNGCFLAGS} -c -o $@ $<
-src/gfx/proto_palette.o: src/gfx/proto_palette.cpp
+src/gfx/process.o: src/gfx/process.cpp
 	$Q${CXX} ${REALCXXFLAGS} ${PNGCFLAGS} -c -o $@ $<
 src/gfx/reverse.o: src/gfx/reverse.cpp
 	$Q${CXX} ${REALCXXFLAGS} ${PNGCFLAGS} -c -o $@ $<
@@ -163,7 +175,6 @@ src/gfx/rgba.o: src/gfx/rgba.cpp
 	$Q${CXX} ${REALCXXFLAGS} -c -o $@ $<
 
 # Target used to remove all files generated by other Makefile targets
-
 clean:
 	$Q${RM} rgbasm rgbasm.exe
 	$Q${RM} rgblink rgblink.exe
@@ -177,7 +188,6 @@ clean:
 	$Q${RM} test/gfx/randtilegen test/gfx/rgbgfx_test
 
 # Target used to install the binaries and man pages.
-
 install: all
 	$Qinstall -d ${DESTDIR}${bindir}/ ${DESTDIR}${mandir}/man1/ ${DESTDIR}${mandir}/man5/ ${DESTDIR}${mandir}/man7/
 	$Qinstall ${STRIP} -m ${BINMODE} rgbasm ${DESTDIR}${bindir}/rgbasm${SUFFIX}
@@ -189,15 +199,13 @@ install: all
 	$Qinstall -m ${MANMODE} man/rgbds.7 man/gbz80.7 ${DESTDIR}${mandir}/man7/
 
 # Target used to check for suspiciously missing changed files.
-
 checkdiff:
 	$Qcontrib/checkdiff.bash `git merge-base HEAD ${BASE_REF}`
 
-# This target is used during development in order to prevent adding new issues
-# to the source code. All warnings are treated as errors in order to block the
-# compilation and make the continous integration infrastructure return failure.
+# Target used in development to prevent adding new issues to the source code.
+# All warnings are treated as errors to block the compilation and make the
+# continous integration infrastructure return failure.
 # The rationale for some of the flags is documented in the CMakeLists.
-
 develop:
 	$Q${MAKE} WARNFLAGS="${WARNFLAGS} -Werror -Wextra \
 		-Walloc-zero -Wcast-align -Wcast-qual -Wduplicated-branches -Wduplicated-cond \
@@ -210,26 +218,27 @@ develop:
 		-fsanitize=float-divide-by-zero" \
 		CXXFLAGS="-ggdb3 -Og -fno-omit-frame-pointer -fno-optimize-sibling-calls"
 
-# This target is used during development in order to more easily debug with gdb.
-
+# Target used in development to debug with gdb.
 debug:
 	$Qenv ${MAKE} \
 		CXXFLAGS="-ggdb3 -O0 -fno-omit-frame-pointer -fno-optimize-sibling-calls"
 
-# This target is used during development in order to more easily profile with callgrind.
-
+# Target used in development to profile with callgrind.
 profile:
 	$Qenv ${MAKE} \
 		CXXFLAGS="-ggdb3 -O3 -fno-omit-frame-pointer -fno-optimize-sibling-calls"
 
-# This target is used during development in order to inspect code coverage with gcov.
-
+# Target used in development to inspect code coverage with gcov.
 coverage:
 	$Qenv ${MAKE} \
 		CXXFLAGS="-ggdb3 -Og --coverage -fno-omit-frame-pointer -fno-optimize-sibling-calls"
 
-# This target is used during development in order to remove unused `#include` headers.
+# Target used in development to check code with clang-tidy.
+# Requires Bison-generated header files to exist.
+tidy: src/asm/parser.hpp src/link/script.hpp
+	$Qclang-tidy -p . $$(find src -name '*.cpp')
 
+# Target used in development to remove unused `#include` headers.
 iwyu:
 	$Qenv ${MAKE} \
 		CXX="include-what-you-use" \

README.md

@@ -6,7 +6,7 @@ for the Game Boy and Game Boy Color. It consists of:
 - RGBASM (assembler)
 - RGBLINK (linker)
 - RGBFIX (checksum/header fixer)
-- RGBGFX (PNG‐to‐Game Boy graphics converter)
+- RGBGFX (PNG-to-Game Boy graphics converter)
 
 This is a fork of the original RGBDS which aims to make the programs more like
 other UNIX tools.
@@ -93,6 +93,7 @@ The RGBDS source code file structure is as follows:
 │   └── run-tests.sh
 ├── .clang-format
 ├── CMakeLists.txt
+├── compile_flags.txt
 ├── Dockerfile
 ├── Makefile
 └── README.md
@@ -118,6 +119,8 @@ The RGBDS source code file structure is as follows:
   modify the behavior of RGBDS.
 - `.clang-format` - code style for automated C++ formatting with
   [`clang-format`](https://clang.llvm.org/docs/ClangFormat.html).
+- `compile_flags.txt` - compiler flags for C++ static analysis with
+  [`clang-tidy`](https://clang.llvm.org/extra/clang-tidy/).
 - `Dockerfile` - defines how to build RGBDS with Docker.
 
 ## 3. History
@@ -140,7 +143,7 @@ The RGBDS source code file structure is as follows:
 - 2010-09-25: Sørensen continues development of
   [ASMotor](https://github.com/asmotor/asmotor) to this day.
 - 2015-01-18: stag019 begins implementing [RGBGFX](https://github.com/stag019/rgbgfx),
-  a PNG‐to‐Game Boy graphics converter, for eventual integration into RGBDS.
+  a PNG-to-Game Boy graphics converter, for eventual integration into RGBDS.
 - 2016-09-05: RGBGFX is
   [integrated](https://github.com/gbdev/rgbds/commit/c3c31138ddbd8680d4e67957e387f2816798a71b)
   into Bentley's repository.

RELEASE.md

@@ -68,7 +68,7 @@ GitHub.
 
 6. Click the "Publish release" button to publish it!
 
-7. Update the `release` branch. You can use `git push origin release`.
+7. Update the `release` branch. You can use `git push origin master:release`.
 
 8. Update the following related projects.
 
@@ -77,9 +77,7 @@ GitHub.
       to list the new release.
    2. [rgbds-live](https://github.com/gbdev/rgbds-live): update the `rgbds` submodule (and
       [patches/rgbds.patch](https://github.com/gbdev/rgbds-live/blob/master/patches/rgbds.patch)
-      if necessary) to use the new release, and
-      [index.html](https://github.com/gbdev/rgbds-live/blob/master/index.html)
-      to link to the new manual version.
+      if necessary) to use the new release.
    3. [rgbobj](https://github.com/gbdev/rgbobj) and [rgbds-obj](https://github.com/gbdev/rgbds-obj):
        make sure that object files created by the latest RGBASM can be parsed and displayed.
        If the object file revision has been updated, `rgbobj` will need a corresponding release.

compile_flags.txt

@@ -0,0 +1,6 @@
+-std=c++2a
+-I
+include
+-fno-exceptions
+-fno-rtti
+-fno-caret-diagnostics

contrib/bash_compl/_rgbasm.bash

@@ -24,13 +24,13 @@ _rgbasm_completions() {
 	# Empty long opt = it doesn't exit
 	# See the `state` variable below for info about `state_after`
 	declare -A opts=(
-		[V]="version:normal"
 		[h]="help:normal"
-		[E]="export-all:normal"
-		[v]="verbose:normal"
+		[V]="version:normal"
+		[W]="warning:warning"
 		[w]=":normal"
 		[b]="binary-digits:unk"
 		[D]="define:unk"
+		[E]="export-all:normal"
 		[g]="gfx-chars:unk"
 		[I]="include:dir"
 		[M]="dependfile:glob-*.mk *.d"
@@ -40,7 +40,7 @@ _rgbasm_completions() {
 		[Q]="q-precision:unk"
 		[r]="recursion-depth:unk"
 		[s]="state:unk"
-		[W]="warning:warning"
+		[v]="verbose:normal"
 		[X]="max-errors:unk"
 	)
 	# Parse command-line up to current word
@@ -155,7 +155,7 @@ _rgbasm_completions() {
 			parse_short_opt "$cur_word"
 
 			if [[ "$state" = 'normal' ]]; then
-				mapfile -t COMPREPLY < <(compgen -W "${!opts[*]}" -P "$cur_word" ''; compgen -W '-MG -MP -MQ -MT' "$cur_word")
+				mapfile -t COMPREPLY < <(compgen -W "${!opts[*]}" -P "$cur_word" ''; compgen -W '-MC -MG -MP -MQ -MT' "$cur_word")
 				return 0
 			elif [[ "$optlen" = "${#cur_word}" && "$state" != "warning" ]]; then
 				# This short option group only awaits its argument!

contrib/bash_compl/_rgbfix.bash

@@ -7,15 +7,15 @@ _rgbfix_completions() {
 	# Empty long opt = it doesn't exit
 	# See the `state` variable below for info about `state_after`
 	declare -A opts=(
-		[V]="version:normal"
 		[h]="help:normal"
-		[j]="non-japanese:normal"
-		[s]="sgb-compatible:normal"
-		[v]="validate:normal"
+		[V]="version:normal"
+		[W]="warning:warning"
+		[w]=":normal"
 		[C]="color-only:normal"
 		[c]="color-compatible:normal"
 		[f]="fix-spec:fix-spec"
 		[i]="game-id:unk"
+		[j]="non-japanese:normal"
 		[k]="new-licensee:unk"
 		[L]="custom-logo:glob-*.1bpp"
 		[l]="old-licensee:unk"
@@ -24,7 +24,9 @@ _rgbfix_completions() {
 		[o]="output:glob-*.gb *.gbc *.sgb"
 		[p]="pad-value:unk"
 		[r]="ram-size:unk"
+		[s]="sgb-compatible:normal"
 		[t]="title:unk"
+		[v]="validate:normal"
 	)
 	# Parse command-line up to current word
 	local opt_ena=true
@@ -140,6 +142,16 @@ _rgbfix_completions() {
 	case "$state" in
 		unk) # Return with no replies: no idea what to complete!
 			;;
+		warning)
+			mapfile -t COMPREPLY < <(compgen -W "
+				mbc
+				overwrite
+				sgb
+				truncation
+				all
+				everything
+				error" -P "${cur_word:0:$optlen}" -- "${cur_word:$optlen}")
+			;;
 		fix-spec)
 			COMPREPLY=( "${cur_word}"{l,h,g,L,H,G} )
 			;;

contrib/bash_compl/_rgbgfx.bash

@@ -7,36 +7,38 @@ _rgbgfx_completions() {
 	# Empty long opt = it doesn't exit
 	# See the `state` variable below for info about `state_after`
 	declare -A opts=(
-		[V]="version:normal"
 		[h]="help:normal"
-		[C]="color-curve:normal"
-		[m]="mirror-tiles:normal"
-		[O]="group-outputs:normal"
-		[u]="unique-tiles:normal"
-		[v]="verbose:normal"
-		[X]="mirror-x:normal"
-		[Y]="mirror-y:normal"
-		[Z]="columns:normal"
-		[a]="attr-map:glob-*.attrmap"
+		[V]="version:normal"
+		[W]="warning:warning"
+		[w]=":normal"
 		[A]="auto-attr-map:normal"
+		[a]="attr-map:glob-*.attrmap"
 		[B]="background-color:unk"
 		[b]="base-tiles:unk"
+		[C]="color-curve:normal"
 		[c]="colors:unk"
 		[d]="depth:unk"
 		[i]="input-tileset:glob-*.2bpp"
 		[L]="slice:unk"
+		[m]="mirror-tiles:normal"
 		[N]="nb-tiles:unk"
 		[n]="nb-palettes:unk"
+		[O]="group-outputs:normal"
 		[o]="output:glob-*.2bpp"
-		[p]="palette:glob-*.pal"
 		[P]="auto-palette:normal"
-		[q]="palette-map:glob-*.palmap"
+		[p]="palette:glob-*.pal"
 		[Q]="auto-palette-map:normal"
+		[q]="palette-map:glob-*.palmap"
 		[r]="reverse:unk"
 		[s]="palette-size:unk"
-		[t]="tilemap:glob-*.tilemap"
 		[T]="auto-tilemap:normal"
+		[t]="tilemap:glob-*.tilemap"
+		[u]="unique-tiles:normal"
+		[v]="verbose:normal"
+		[X]="mirror-x:normal"
 		[x]="trim-end:unk"
+		[Y]="mirror-y:normal"
+		[Z]="columns:normal"
 	)
 	# Parse command-line up to current word
 	local opt_ena=true
@@ -152,6 +154,14 @@ _rgbgfx_completions() {
 	case "$state" in
 		unk) # Return with no replies: no idea what to complete!
 			;;
+		warning)
+			mapfile -t COMPREPLY < <(compgen -W "
+				embedded
+				trim-nonempty
+				all
+				everything
+				error" -P "${cur_word:0:$optlen}" -- "${cur_word:$optlen}")
+			;;
 		normal) # Acts like a glob...
 			state="glob-*.png"
 			;&

contrib/bash_compl/_rgblink.bash

@@ -7,20 +7,21 @@ _rgblink_completions() {
 	# Empty long opt = it doesn't exit
 	# See the `state` variable below for info about `state_after`
 	declare -A opts=(
-		[V]="version:normal"
 		[h]="help:normal"
-		[d]="dmg:normal"
-		[t]="tiny:normal"
-		[v]="verbose:normal"
-		[w]="wramx:normal"
-		[x]="nopad:normal"
-		[l]="linkerscript:glob-*"
+		[V]="version:normal"
+		[W]="warning:warning"
 		[M]="no-sym-in-map:normal"
+		[d]="dmg:normal"
+		[l]="linkerscript:glob-*"
 		[m]="map:glob-*.map"
 		[n]="sym:glob-*.sym"
 		[O]="overlay:glob-*.gb *.gbc *.sgb"
 		[o]="output:glob-*.gb *.gbc *.sgb"
 		[p]="pad:unk"
+		[t]="tiny:normal"
+		[v]="verbose:normal"
+		[w]="wramx:normal"
+		[x]="nopad:normal"
 	)
 	# Parse command-line up to current word
 	local opt_ena=true
@@ -136,6 +137,18 @@ _rgblink_completions() {
 	case "$state" in
 		unk) # Return with no replies: no idea what to complete!
 			;;
+		warning)
+			mapfile -t COMPREPLY < <(compgen -W "
+				assert
+				div
+				obsolete
+				shift
+				shift-amount
+				truncation
+				all
+				everything
+				error" -P "${cur_word:0:$optlen}" -- "${cur_word:$optlen}")
+			;;
 		normal) # Acts like a glob...
 			state="glob-*.o *.obj"
 			;&

contrib/checkdiff.bash

@@ -26,11 +26,23 @@ dependency include/linkdefs.hpp    man/rgbds.5 \
 dependency src/asm/parser.y        man/rgbasm.5 \
            "Was the rgbasm grammar changed?"
 
+dependency src/asm/actions.cpp     man/rgbasm.5 \
+           "Was the rgbasm grammar changed?"
+
 dependency src/link/script.y       man/rgblink.5 \
            "Was the linker script grammar changed?"
 
+dependency src/link/layout.cpp     man/rgblink.5 \
+           "Was the linker script grammar changed?"
+
 dependency include/asm/warning.hpp  man/rgbasm.1 \
            "Were the rgbasm warnings changed?"
+dependency include/link/warning.hpp man/rgblink.1 \
+           "Were the rgblink warnings changed?"
+dependency include/fix/warning.hpp  man/rgbfix.1 \
+           "Were the rgbfix warnings changed?"
+dependency include/gfx/warning.hpp  man/rgbgfx.1 \
+           "Were the rgbgfx warnings changed?"
 
 dependency src/asm/object.cpp      include/linkdefs.hpp \
            "Should the object file revision be bumped?"

contrib/checkformat.bash

@@ -11,5 +11,5 @@ if ! git diff-index --quiet HEAD --; then
 	git diff-index --name-only HEAD --
 	echo
 	git diff HEAD --
-	return 1
+	exit 1
 fi

contrib/coverage.bash

@@ -7,20 +7,29 @@ make coverage -j
 # Run the tests
 pushd test
 ./fetch-test-deps.sh
-./run-tests.sh
+if [[ $# -eq 0 ]]; then
+  ./run-tests.sh
+else
+  ./run-tests.sh --os "$1"
+fi
 popd
 
 # Generate coverage logs
 gcov src/**/*.cpp
 mkdir -p coverage
 
-# Generate coverage report
-lcov -c --no-external -d . -o coverage/coverage.info
-genhtml -f -s -o coverage/ coverage/coverage.info
+# Generate coverage report, excluding Bison-generated files
+COVERAGE_INFO=coverage/coverage.info
+lcov -c --no-external -d . -o "$COVERAGE_INFO"
+lcov -r "$COVERAGE_INFO" src/asm/parser.{hpp,cpp} src/link/script.{hpp,cpp} -o "$COVERAGE_INFO"
+genhtml --dark-mode -f -s -o coverage/ "$COVERAGE_INFO"
 
-# Open report in web browser
-if [ "$(uname)" == "Darwin" ]; then
+# Check whether running from coverage.yml workflow
+if [ "$1" != "ubuntu-ci" ]; then
+  # Open report in web browser
+  if [ "$(uname)" == "Darwin" ]; then
     open coverage/index.html
-else
+  else
     xdg-open coverage/index.html
+  fi
 fi

contrib/zsh_compl/_rgbasm

@@ -23,15 +23,13 @@ _rgbasm_warnings() {
 		'obsolete:Warn when using deprecated features'
 		'purge:Warn when purging exported symbols or labels'
 		'shift:Warn when shifting negative values'
-		'shift-amount:Warn when a shift'\''s operand it negative or \> 32'
+		'shift-amount:Warn when a shift'\''s operand is negative or \> 32'
 		'truncation:Warn when implicit truncation loses bits'
 		'unmapped-char:Warn on unmapped character'
 		'unmatched-directive:Warn on unmatched directive pair'
 		'unterminated-load:Warn on LOAD without ENDL'
 		'user:Warn when executing the WARN built-in'
 	)
-	# TODO: handle `no-` and `error=` somehow?
-	# TODO: handle `=0|1|2` levels for `numeric-string`, `purge`, `truncation`, and `unmapped-char`?
 	_describe warning warnings
 }
 
@@ -48,9 +46,10 @@ local args=(
 	'*'{-D,--define}'+[Define a string symbol]:name + value (default 1):'
 	'(-g --gfx-chars)'{-g,--gfx-chars}'+[Change chars for gfx constants]:chars spec:'
 	'(-I --include)'{-I,--include}'+[Add an include directory]:include path:_files -/'
-	'(-M --dependfile)'{-M,--dependfile}"+[Write deps in make format]:output file:_files -g '*.{d,mk}'"
-	-MG'[Assume missing files should be generated]'
-	-MP'[Add phony targets to all deps]'
+	'(-M --dependfile)'{-M,--dependfile}"+[Write dependencies in Makefile format]:output file:_files -g '*.{d,mk}'"
+	-MC'[Continue after missing dependencies]'
+	-MG'[Assume missing dependencies should be generated]'
+	-MP'[Add phony targets to all dependencies]'
 	'*'-MT"+[Add a target to the rules]:target:_files -g '*.{d,mk,o}'"
 	'*'-MQ"+[Add a target to the rules]:target:_files -g '*.{d,mk,o}'"
 	'(-o --output)'{-o,--output}'+[Output file]:output file:_files'

contrib/zsh_compl/_rgbfix

@@ -34,6 +34,21 @@ _mbc_names() {
 	_describe "MBC name" mbc_names
 }
 
+_rgbfix_warnings() {
+	local warnings=(
+		'error:Turn all warnings into errors'
+
+		'all:Enable most warning messages'
+		'everything:Enable literally everything'
+
+		'mbc:Warn about issues with MBC specs'
+		'overwrite:Warn when overwriting non-zero bytes'
+		'sgb:Warn when SGB flag conflicts with old licensee code'
+		'truncation:Warn when values are truncated to fit'
+	)
+	_describe warning warnings
+}
+
 local args=(
 	# Arguments are listed here in the same order as in the manual, except for the version and help
 	'(- : * options)'{-V,--version}'[Print version number and exit]'
@@ -45,6 +60,7 @@ local args=(
 	'(-O --overwrite)'{-O,--overwrite}'[Allow overwriting non-zero bytes]'
 	'(-s --sgb-compatible)'{-s,--sgb-compatible}'[Set the SGB flag]'
 	'(-f --fix-spec -v --validate)'{-v,--validate}'[Shorthand for -f lhg]'
+	-w'[Disable all warnings]'
 
 	'(-f --fix-spec -v --validate)'{-f,--fix-spec}'+[Fix or trash some header values]:fix spec:'
 	'(-i --game-id)'{-i,--game-id}'+[Set game ID string]:4-char game ID:'
@@ -57,6 +73,7 @@ local args=(
 	'(-p --pad-value)'{-p,--pad-value}'+[Pad to next valid size using this byte as padding]:padding byte:'
 	'(-r --ram-size)'{-r,--ram-size}'+[Set RAM size]:ram size byte:'
 	'(-t --title)'{-t,--title}'+[Set title string]:11-char title string:'
+	'(-W --warning)'{-W,--warning}'+[Toggle warning flags]:warning flag:_rgbfix_warnings'
 
 	'*'":ROM files:_files -g '*.{gb,sgb,gbc}'"
 )

contrib/zsh_compl/_rgbgfx

@@ -9,6 +9,19 @@ _depths() {
 	_describe 'bit depth' depths
 }
 
+_rgbgfx_warnings() {
+	local warnings=(
+		'error:Turn all warnings into errors'
+
+		'all:Enable most warning messages'
+		'everything:Enable literally everything'
+
+		'embedded:Warn when using embedded PLTE without "-c embedded"'
+		'trim-nonempty:Warn when "-x" trims nonempty tiles'
+	)
+	_describe warning warnings
+}
+
 local args=(
 	# Arguments are listed here in the same order as in the manual, except for the version and help
 	'(- : * options)'{-V,--version}'[Print version number and exit]'
@@ -22,7 +35,8 @@ local args=(
 	'(-q --palette-map -Q --auto-palette-map)'{-Q,--auto-palette-map}'[Shortcut for -p <file>.palmap]'
 	'(-t --tilemap -T --auto-tilemap)'{-T,--auto-tilemap}'[Shortcut for -t <file>.tilemap]'
 	'(-u --unique-tiles)'{-u,--unique-tiles}'[Eliminate redundant tiles]'
-	{-v,--verbose}'[Enable verbose output]'
+	'(-v --verbose)'{-v,--verbose}'[Enable verbose output]'
+	-w'[Disable all warnings]'
 	'(-X --mirror-x)'{-X,--mirror-x}'[Eliminate horizontally mirrored tiles from output]'
 	'(-Y --mirror-y)'{-Y,--mirror-y}'[Eliminate vertically mirrored tiles from output]'
 	'(-Z --columns)'{-Z,--columns}'[Read the image in column-major order]'
@@ -42,6 +56,7 @@ local args=(
 	'(-r --reverse)'{-r,--reverse}'+[Yield an image from binary data]:image width (in tiles):'
 	'(-s --palette-size)'{-s,--palette-size}'+[Limit palette size]:palette size:'
 	'(-t --tilemap -T --auto-tilemap)'{-t,--tilemap}'+[Generate a map of tile indices]:tilemap file:_files'
+	'(-W --warning)'{-W,--warning}'+[Toggle warning flags]:warning flag:_rgbgfx_warnings'
 	'(-x --trim-end)'{-x,--trim-end}'+[Trim end of output by this many tiles]:tile count:'
 
 	":input png file:_files -g '*.png'"

contrib/zsh_compl/_rgblink

@@ -1,5 +1,22 @@
 #compdef rgblink
 
+_rgblink_warnings() {
+	local warnings=(
+		'error:Turn all warnings into errors'
+
+		'all:Enable most warning messages'
+		'everything:Enable literally everything'
+
+		'assert:Warn when WARN-type asserts fail'
+		'div:Warn when dividing the smallest int by -1'
+		'obsolete:Warn when using deprecated features'
+		'shift:Warn when shifting negative values'
+		'shift-amount:Warn when a shift'\''s operand is negative or \> 32'
+		'truncation:Warn when implicit truncation loses bits'
+	)
+	_describe warning warnings
+}
+
 local args=(
 	# Arguments are listed here in the same order as in the manual, except for the version and help
 	'(- : * options)'{-V,--version}'[Print version number and exit]'
@@ -19,6 +36,7 @@ local args=(
 	'(-o --output)'{-o,--output}"+[Write ROM image to this file]:rom file:_files -g '*.{gb,sgb,gbc}'"
 	'(-p --pad-value)'{-p,--pad-value}'+[Set padding byte]:padding byte:'
 	'(-S --scramble)'{-s,--scramble}'+[Activate scrambling]:scramble spec'
+	'(-W --warning)'{-W,--warning}'+[Toggle warning flags]:warning flag:_rgblink_warnings'
 
 	'*'":object files:_files -g '*.o'"
 )

include/asm/actions.hpp

@@ -0,0 +1,42 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_ASM_ACTIONS_HPP
+#define RGBDS_ASM_ACTIONS_HPP
+
+#include <optional>
+#include <stdint.h>
+#include <string>
+#include <string_view>
+#include <variant>
+#include <vector>
+
+#include "asm/output.hpp" // AssertionType
+#include "asm/rpn.hpp"    // RPNCommand
+
+std::optional<std::string> act_ReadFile(std::string const &name, uint32_t maxLen);
+
+uint32_t act_StringToNum(std::vector<int32_t> const &str);
+
+size_t act_StringLen(std::string const &str, bool printErrors);
+std::string act_StringSlice(std::string const &str, uint32_t start, uint32_t stop);
+std::string act_StringSub(std::string const &str, uint32_t pos, uint32_t len);
+
+size_t act_CharLen(std::string const &str);
+std::string act_StringChar(std::string const &str, uint32_t idx);
+std::string act_CharSub(std::string const &str, uint32_t pos);
+int32_t act_CharCmp(std::string_view str1, std::string_view str2);
+
+uint32_t act_AdjustNegativeIndex(int32_t idx, size_t len, char const *functionName);
+uint32_t act_AdjustNegativePos(int32_t pos, size_t len, char const *functionName);
+
+std::string act_StringReplace(std::string_view str, std::string const &old, std::string const &rep);
+std::string act_StringFormat(
+    std::string const &spec, std::vector<std::variant<uint32_t, std::string>> const &args
+);
+
+void act_CompoundAssignment(std::string const &symName, RPNCommand op, int32_t constValue);
+
+void act_FailAssert(AssertionType type);
+void act_FailAssertMsg(AssertionType type, std::string const &message);
+
+#endif // RGBDS_ASM_ACTIONS_HPP

include/asm/fixpoint.hpp

@@ -5,9 +5,6 @@
 
 #include <stdint.h>
 
-extern uint8_t fixPrecision;
-
-uint8_t fix_Precision();
 int32_t fix_Sin(int32_t i, int32_t q);
 int32_t fix_Cos(int32_t i, int32_t q);
 int32_t fix_Tan(int32_t i, int32_t q);

include/asm/fstack.hpp

@@ -10,16 +10,16 @@
 #include <stdint.h>
 #include <stdio.h>
 #include <string>
+#include <variant>
 #include <vector>
 
-#include "either.hpp"
 #include "linkdefs.hpp"
 
 #include "asm/lexer.hpp"
 
 struct FileStackNode {
 	FileStackNodeType type;
-	Either<
+	std::variant<
 	    std::vector<uint32_t>, // NODE_REPT
 	    std::string            // NODE_FILE, NODE_MACRO
 	    >
@@ -34,33 +34,34 @@ struct FileStackNode {
 	uint32_t ID = UINT32_MAX;
 
 	// REPT iteration counts since last named node, in reverse depth order
-	std::vector<uint32_t> &iters() { return data.get<std::vector<uint32_t>>(); }
-	std::vector<uint32_t> const &iters() const { return data.get<std::vector<uint32_t>>(); }
+	std::vector<uint32_t> &iters() { return std::get<std::vector<uint32_t>>(data); }
+	std::vector<uint32_t> const &iters() const { return std::get<std::vector<uint32_t>>(data); }
 	// File name for files, file::macro name for macros
-	std::string &name() { return data.get<std::string>(); }
-	std::string const &name() const { return data.get<std::string>(); }
+	std::string &name() { return std::get<std::string>(data); }
+	std::string const &name() const { return std::get<std::string>(data); }
 
-	FileStackNode(FileStackNodeType type_, Either<std::vector<uint32_t>, std::string> data_)
+	FileStackNode(FileStackNodeType type_, std::variant<std::vector<uint32_t>, std::string> data_)
 	    : type(type_), data(data_) {}
 
 	std::string const &dump(uint32_t curLineNo) const;
+	std::string reptChain() const;
 };
 
-extern size_t maxRecursionDepth;
-
 struct MacroArgs;
 
-void fstk_DumpCurrent();
+bool fstk_DumpCurrent();
 std::shared_ptr<FileStackNode> fstk_GetFileStack();
 std::shared_ptr<std::string> fstk_GetUniqueIDStr();
 MacroArgs *fstk_GetCurrentMacroArgs();
 
 void fstk_AddIncludePath(std::string const &path);
-void fstk_SetPreIncludeFile(std::string const &path);
+void fstk_AddPreIncludeFile(std::string const &path);
 std::optional<std::string> fstk_FindFile(std::string const &path);
+bool fstk_FileError(std::string const &path, char const *functionName);
+bool fstk_FailedOnMissingInclude();
 
 bool yywrap();
-void fstk_RunInclude(std::string const &path, bool updateStateNow);
+bool fstk_RunInclude(std::string const &path);
 void fstk_RunMacro(std::string const &macroName, std::shared_ptr<MacroArgs> macroArgs);
 void fstk_RunRept(uint32_t count, int32_t reptLineNo, ContentSpan const &span);
 void fstk_RunFor(
@@ -74,6 +75,6 @@ void fstk_RunFor(
 bool fstk_Break();
 
 void fstk_NewRecursionDepth(size_t newDepth);
-void fstk_Init(std::string const &mainPath, size_t maxDepth);
+void fstk_Init(std::string const &mainPath);
 
 #endif // RGBDS_ASM_FSTACK_HPP

include/asm/lexer.hpp

@@ -8,13 +8,13 @@
 #include <optional>
 #include <stdint.h>
 #include <string>
+#include <variant>
 #include <vector>
 
-#include "either.hpp"
 #include "platform.hpp" // SSIZE_MAX
 
-// This value is a compromise between `LexerState` allocation performance when `mmap` works, and
-// buffering performance when it doesn't/can't (e.g. when piping a file into RGBASM).
+// This value is a compromise between `LexerState` allocation performance when reading the entire
+// file works, and buffering performance when it doesn't (e.g. when piping a file into RGBASM).
 static constexpr size_t LEXER_BUF_SIZE = 64;
 // The buffer needs to be large enough for the maximum `lexerState->peek()` lookahead distance
 static_assert(LEXER_BUF_SIZE > 1, "Lexer buffer size is too small");
@@ -84,6 +84,7 @@ struct LexerState {
 	bool atLineStart;
 	uint32_t lineNo;
 	int lastToken;
+	int nextToken;
 
 	std::deque<IfStackEntry> ifStack;
 
@@ -91,13 +92,12 @@ struct LexerState {
 	size_t captureSize; // Amount of text captured
 	std::shared_ptr<std::vector<char>> captureBuf; // Buffer to send the captured text to if set
 
-	bool disableMacroArgs;
-	bool disableInterpolation;
-	size_t macroArgScanDistance; // Max distance already scanned for macro args
+	bool disableExpansions;
+	size_t expansionScanDistance; // Max distance already scanned for expansions
 	bool expandStrings;
 	std::deque<Expansion> expansions; // Front is the innermost current expansion
 
-	Either<ViewedContent, BufferedContent> content;
+	std::variant<std::monostate, ViewedContent, BufferedContent> content;
 
 	~LexerState();
 
@@ -109,21 +109,17 @@ struct LexerState {
 	}
 
 	void setAsCurrentState();
-	bool setFileAsNextState(std::string const &filePath, bool updateStateNow);
+	void setFileAsNextState(std::string const &filePath, bool updateStateNow);
 	void setViewAsNextState(char const *name, ContentSpan const &span, uint32_t lineNo_);
 
 	void clear(uint32_t lineNo_);
 };
 
-extern char binDigits[2];
-extern char gfxDigits[4];
-
 void lexer_SetBinDigits(char const digits[2]);
 void lexer_SetGfxDigits(char const digits[4]);
 
 bool lexer_AtTopLevel();
 void lexer_RestartRept(uint32_t lineNo);
-void lexer_Init();
 void lexer_SetMode(LexerMode mode);
 void lexer_ToggleStringExpansion(bool enable);
 

include/asm/main.hpp

@@ -3,16 +3,50 @@
 #ifndef RGBDS_ASM_MAIN_HPP
 #define RGBDS_ASM_MAIN_HPP
 
+#include <stdint.h>
 #include <stdio.h>
 #include <string>
 
-extern bool verbose;
-extern bool warnings; // True to enable warnings, false to disable them.
+enum MissingInclude {
+	INC_ERROR,    // A missing included file is an error that halts assembly
+	GEN_EXIT,     // A missing included file is assumed to be generated; exit normally
+	GEN_CONTINUE, // A missing included file is assumed to be generated; continue assembling
+};
 
-extern FILE *dependFile;
-extern std::string targetFileName;
-extern bool generatedMissingIncludes;
-extern bool failedOnMissingInclude;
-extern bool generatePhonyDeps;
+struct Options {
+	uint8_t fixPrecision = 16;                      // -Q
+	size_t maxRecursionDepth = 64;                  // -r
+	char binDigits[2] = {'0', '1'};                 // -b
+	char gfxDigits[4] = {'0', '1', '2', '3'};       // -g
+	bool verbose = false;                           // -v
+	FILE *dependFile = nullptr;                     // -M
+	std::string targetFileName;                     // -MQ, -MT
+	MissingInclude missingIncludeState = INC_ERROR; // -MC, -MG
+	bool generatePhonyDeps = false;                 // -MP
+	std::string objectFileName;                     // -o
+	uint8_t padByte = 0;                            // -p
+	uint64_t maxErrors = 0;                         // -X
+
+	~Options() {
+		if (dependFile) {
+			fclose(dependFile);
+		}
+	}
+
+	void printDep(std::string const &depName) {
+		if (dependFile) {
+			fprintf(dependFile, "%s: %s\n", targetFileName.c_str(), depName.c_str());
+		}
+	}
+};
+
+extern Options options;
+
+#define verbosePrint(...) \
+	do { \
+		if (options.verbose) { \
+			fprintf(stderr, __VA_ARGS__); \
+		} \
+	} while (0)
 
 #endif // RGBDS_ASM_MAIN_HPP

include/asm/opt.hpp

@@ -5,10 +5,10 @@
 
 #include <stdint.h>
 
-void opt_B(char const chars[2]);
-void opt_G(char const chars[4]);
+void opt_B(char const binDigits[2]);
+void opt_G(char const gfxDigits[4]);
 void opt_P(uint8_t padByte);
-void opt_Q(uint8_t precision);
+void opt_Q(uint8_t fixPrecision);
 void opt_W(char const *flag);
 void opt_Parse(char const *option);
 

include/asm/output.hpp

@@ -16,10 +16,7 @@ struct FileStackNode;
 
 enum StateFeature { STATE_EQU, STATE_VAR, STATE_EQUS, STATE_CHAR, STATE_MACRO, NB_STATE_FEATURES };
 
-extern std::string objectFileName;
-
 void out_RegisterNode(std::shared_ptr<FileStackNode> node);
-void out_SetFileName(std::string const &name);
 void out_CreatePatch(uint32_t type, Expression const &expr, uint32_t ofs, uint32_t pcShift);
 void out_CreateAssert(
     AssertionType type, Expression const &expr, std::string const &message, uint32_t ofs

include/asm/rpn.hpp

@@ -5,15 +5,15 @@
 
 #include <stdint.h>
 #include <string>
+#include <variant>
 #include <vector>
 
-#include "either.hpp"
 #include "linkdefs.hpp"
 
 struct Symbol;
 
 struct Expression {
-	Either<
+	std::variant<
 	    int32_t,    // If the expression's value is known, it's here
 	    std::string // Why the expression is not known, if it isn't
 	    >
@@ -22,8 +22,8 @@ struct Expression {
 	std::vector<uint8_t> rpn{}; // Bytes serializing the RPN expression
 	uint32_t rpnPatchSize = 0;  // Size the expression will take in the object file
 
-	bool isKnown() const { return data.holds<int32_t>(); }
-	int32_t value() const { return data.get<int32_t>(); }
+	bool isKnown() const { return std::holds_alternative<int32_t>(data); }
+	int32_t value() const { return std::get<int32_t>(data); }
 
 	int32_t getConstVal() const;
 	Symbol const *symbolOf() const;

include/asm/section.hpp

@@ -5,6 +5,7 @@
 
 #include <deque>
 #include <memory>
+#include <optional>
 #include <stdint.h>
 #include <string>
 #include <unordered_map>
@@ -12,8 +13,6 @@
 
 #include "linkdefs.hpp"
 
-extern uint8_t fillByte;
-
 struct Expression;
 struct FileStackNode;
 struct Section;
@@ -42,6 +41,7 @@ struct Section {
 	std::deque<Patch> patches;
 	std::vector<uint8_t> data;
 
+	uint32_t getID() const; // ID of the section in the object file (`UINT32_MAX` if none)
 	bool isSizeKnown() const;
 };
 
@@ -51,9 +51,8 @@ struct SectionSpec {
 	uint16_t alignOfs;
 };
 
-extern std::deque<Section> sectionList;
-extern std::unordered_map<std::string, size_t> sectionMap; // Indexes into `sectionList`
-extern Section *currentSection;
+size_t sect_CountSections();
+void sect_ForEach(void (*callback)(Section &));
 
 Section *sect_FindSectionByName(std::string const &name);
 void sect_NewSection(
@@ -76,6 +75,10 @@ void sect_CheckLoadClosed();
 Section *sect_GetSymbolSection();
 uint32_t sect_GetSymbolOffset();
 uint32_t sect_GetOutputOffset();
+std::optional<uint32_t> sect_GetOutputBank();
+
+Patch *sect_AddOutputPatch();
+
 uint32_t sect_GetAlignBytes(uint8_t alignment, uint16_t offset);
 void sect_AlignPC(uint8_t alignment, uint16_t offset);
 
@@ -87,21 +90,23 @@ void sect_EndUnion();
 void sect_CheckUnionClosed();
 
 void sect_ConstByte(uint8_t byte);
-void sect_ByteString(std::vector<int32_t> const &string);
-void sect_WordString(std::vector<int32_t> const &string);
-void sect_LongString(std::vector<int32_t> const &string);
+void sect_ByteString(std::vector<int32_t> const &str);
+void sect_WordString(std::vector<int32_t> const &str);
+void sect_LongString(std::vector<int32_t> const &str);
 void sect_Skip(uint32_t skip, bool ds);
 void sect_RelByte(Expression const &expr, uint32_t pcShift);
 void sect_RelBytes(uint32_t n, std::vector<Expression> const &exprs);
 void sect_RelWord(Expression const &expr, uint32_t pcShift);
 void sect_RelLong(Expression const &expr, uint32_t pcShift);
 void sect_PCRelByte(Expression const &expr, uint32_t pcShift);
-void sect_BinaryFile(std::string const &name, int32_t startPos);
-void sect_BinaryFileSlice(std::string const &name, int32_t startPos, int32_t length);
+bool sect_BinaryFile(std::string const &name, uint32_t startPos);
+bool sect_BinaryFileSlice(std::string const &name, uint32_t startPos, uint32_t length);
 
 void sect_EndSection();
 void sect_PushSection();
 void sect_PopSection();
 void sect_CheckStack();
 
+std::string sect_PushSectionFragmentLiteral();
+
 #endif // RGBDS_ASM_SECTION_HPP

include/asm/symbol.hpp

@@ -24,8 +24,8 @@ enum SymbolType {
 	SYM_REF // Forward reference to a label
 };
 
-struct Symbol;                    // For the `sym_IsPC` forward declaration
-bool sym_IsPC(Symbol const *sym); // For the inline `getSection` method
+struct Symbol;                    // Forward declaration for `sym_IsPC`
+bool sym_IsPC(Symbol const *sym); // Forward declaration for `getSection`
 
 struct Symbol {
 	std::string name;

include/asm/warning.hpp

@@ -3,7 +3,16 @@
 #ifndef RGBDS_ASM_WARNING_HPP
 #define RGBDS_ASM_WARNING_HPP
 
-extern unsigned int nbErrors, maxErrors;
+#include <functional>
+
+#include "diagnostics.hpp"
+
+enum WarningLevel {
+	LEVEL_DEFAULT,    // Warnings that are enabled by default
+	LEVEL_ALL,        // Warnings that probably indicate an error
+	LEVEL_EXTRA,      // Warnings that are less likely to indicate an error
+	LEVEL_EVERYTHING, // Literally every warning
+};
 
 enum WarningID {
 	WARNING_ASSERT,               // Assertions
@@ -43,24 +52,7 @@ enum WarningID {
 	NB_WARNINGS,
 };
 
-enum WarningAbled { WARNING_DEFAULT, WARNING_ENABLED, WARNING_DISABLED };
-
-struct WarningState {
-	WarningAbled state;
-	WarningAbled error;
-
-	void update(WarningState other);
-};
-
-struct Diagnostics {
-	WarningState flagStates[NB_WARNINGS];
-	WarningState metaStates[NB_WARNINGS];
-};
-
-extern Diagnostics warningStates;
-extern bool warningsAreErrors;
-
-void processWarningFlag(char const *flag);
+extern Diagnostics<WarningLevel, WarningID> warnings;
 
 // Used to warn the user about problems that don't prevent the generation of
 // valid code.
@@ -73,7 +65,7 @@ void warning(WarningID id, char const *fmt, ...);
 // It is also used when the assembler goes into an invalid state (for example,
 // when it fails to allocate memory).
 [[gnu::format(printf, 1, 2), noreturn]]
-void fatalerror(char const *fmt, ...);
+void fatal(char const *fmt, ...);
 
 // Used for errors that make it impossible to assemble correctly, but don't
 // affect the following code. The code will fail to assemble but the user will
@@ -82,4 +74,12 @@ void fatalerror(char const *fmt, ...);
 [[gnu::format(printf, 1, 2)]]
 void error(char const *fmt, ...);
 
+// Used for errors that make it impossible to assemble correctly, but don't
+// affect the following code. The code will fail to assemble but the user will
+// get a list of all errors at the end, making it easier to fix all of them at
+// once.
+void error(std::function<void()> callback);
+
+void requireZeroErrors();
+
 #endif // RGBDS_ASM_WARNING_HPP

include/defaultinitvec.hpp

@@ -1,38 +0,0 @@
-// SPDX-License-Identifier: MIT
-
-#ifndef RGBDS_DEFAULT_INIT_ALLOC_HPP
-#define RGBDS_DEFAULT_INIT_ALLOC_HPP
-
-#include <memory>
-#include <vector>
-
-// Allocator adaptor that interposes construct() calls to convert value-initialization
-// (which is what you get with e.g. `vector::resize`) into default-initialization (which does not
-// zero out non-class types).
-// From
-// https://stackoverflow.com/questions/21028299/is-this-behavior-of-vectorresizesize-type-n-under-c11-and-boost-container/21028912#21028912
-template<typename T, typename A = std::allocator<T>>
-class default_init_allocator : public A {
-	using a_t = std::allocator_traits<A>;
-public:
-	template<typename U>
-	struct rebind {
-		using other = default_init_allocator<U, typename a_t::template rebind_alloc<U>>;
-	};
-
-	using A::A; // Inherit the allocator's constructors
-
-	template<typename U>
-	void construct(U *ptr) noexcept(std::is_nothrow_default_constructible_v<U>) {
-		::new (static_cast<void *>(ptr)) U;
-	}
-	template<typename U, typename... Args>
-	void construct(U *ptr, Args &&...args) {
-		a_t::construct(static_cast<A &>(*this), ptr, std::forward<Args>(args)...);
-	}
-};
-
-template<typename T>
-using DefaultInitVec = std::vector<T, default_init_allocator<T>>;
-
-#endif // RGBDS_DEFAULT_INIT_ALLOC_HPP

include/diagnostics.hpp

@@ -0,0 +1,212 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_DIAGNOSTICS_HPP
+#define RGBDS_DIAGNOSTICS_HPP
+
+#include <inttypes.h>
+#include <optional>
+#include <stdarg.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <string>
+#include <vector>
+
+#include "helpers.hpp"
+#include "itertools.hpp"
+
+[[gnu::format(printf, 1, 2)]]
+void warnx(char const *fmt, ...);
+
+enum WarningAbled { WARNING_DEFAULT, WARNING_ENABLED, WARNING_DISABLED };
+
+struct WarningState {
+	WarningAbled state;
+	WarningAbled error;
+
+	void update(WarningState other);
+};
+
+std::pair<WarningState, std::optional<uint32_t>> getInitialWarningState(std::string &flag);
+
+template<typename L>
+struct WarningFlag {
+	char const *name;
+	L level;
+};
+
+enum WarningBehavior { DISABLED, ENABLED, ERROR };
+
+template<typename W>
+struct ParamWarning {
+	W firstID;
+	W lastID;
+	uint8_t defaultLevel;
+};
+
+template<typename W>
+struct DiagnosticsState {
+	WarningState flagStates[W::NB_WARNINGS];
+	WarningState metaStates[W::NB_WARNINGS];
+	bool warningsEnabled = true;
+	bool warningsAreErrors = false;
+};
+
+template<typename L, typename W>
+struct Diagnostics {
+	std::vector<WarningFlag<L>> metaWarnings;
+	std::vector<WarningFlag<L>> warningFlags;
+	std::vector<ParamWarning<W>> paramWarnings;
+	DiagnosticsState<W> state;
+	uint64_t nbErrors;
+
+	void incrementErrors() {
+		if (nbErrors != UINT64_MAX) {
+			++nbErrors;
+		}
+	}
+
+	WarningBehavior getWarningBehavior(W id) const;
+	std::string processWarningFlag(char const *flag);
+};
+
+template<typename L, typename W>
+WarningBehavior Diagnostics<L, W>::getWarningBehavior(W id) const {
+	// Check if warnings are globally disabled
+	if (!state.warningsEnabled) {
+		return WarningBehavior::DISABLED;
+	}
+
+	// Get the state of this warning flag
+	WarningState const &flagState = state.flagStates[id];
+	WarningState const &metaState = state.metaStates[id];
+
+	// If subsequent checks determine that the warning flag is enabled, this checks whether it has
+	// -Werror without -Wno-error=<flag> or -Wno-error=<meta>, which makes it into an error
+	bool warningIsError = state.warningsAreErrors && flagState.error != WARNING_DISABLED
+	                      && metaState.error != WARNING_DISABLED;
+	WarningBehavior enabledBehavior =
+	    warningIsError ? WarningBehavior::ERROR : WarningBehavior::ENABLED;
+
+	// First, check the state of the specific warning flag
+	if (flagState.state == WARNING_DISABLED) { // -Wno-<flag>
+		return WarningBehavior::DISABLED;
+	}
+	if (flagState.error == WARNING_ENABLED) { // -Werror=<flag>
+		return WarningBehavior::ERROR;
+	}
+	if (flagState.state == WARNING_ENABLED) { // -W<flag>
+		return enabledBehavior;
+	}
+
+	// If no flag is specified, check the state of the "meta" flags that affect this warning flag
+	if (metaState.state == WARNING_DISABLED) { // -Wno-<meta>
+		return WarningBehavior::DISABLED;
+	}
+	if (metaState.error == WARNING_ENABLED) { // -Werror=<meta>
+		return WarningBehavior::ERROR;
+	}
+	if (metaState.state == WARNING_ENABLED) { // -W<meta>
+		return enabledBehavior;
+	}
+
+	// If no meta flag is specified, check the default state of this warning flag
+	if (warningFlags[id].level == L::LEVEL_DEFAULT) { // enabled by default
+		return enabledBehavior;
+	}
+
+	// No flag enables this warning, explicitly or implicitly
+	return WarningBehavior::DISABLED;
+}
+
+template<typename L, typename W>
+std::string Diagnostics<L, W>::processWarningFlag(char const *flag) {
+	std::string rootFlag = flag;
+
+	// Check for `-Werror` or `-Wno-error` to return early
+	if (rootFlag == "error") {
+		// `-Werror` promotes warnings to errors
+		state.warningsAreErrors = true;
+		return rootFlag;
+	} else if (rootFlag == "no-error") {
+		// `-Wno-error` disables promotion of warnings to errors
+		state.warningsAreErrors = false;
+		return rootFlag;
+	}
+
+	auto [flagState, param] = getInitialWarningState(rootFlag);
+
+	// Try to match the flag against a parametric warning
+	// If there was an equals sign, it will have set `param`; if not, `param` will be 0,
+	// which applies to all levels
+	for (ParamWarning<W> const &paramWarning : paramWarnings) {
+		W baseID = paramWarning.firstID;
+		uint8_t maxParam = paramWarning.lastID - baseID + 1;
+		assume(paramWarning.defaultLevel <= maxParam);
+
+		if (rootFlag != warningFlags[baseID].name) {
+			continue;
+		}
+
+		// If making the warning an error but param is 0, set to the maximum
+		// This accommodates `-Werror=<flag>`, but also `-Werror=<flag>=0`, which is
+		// thus filtered out by the caller.
+		// A param of 0 makes sense for disabling everything, but neither for
+		// enabling nor "erroring". Use the default for those.
+		if (!param.has_value() || *param == 0) {
+			param = paramWarning.defaultLevel;
+		} else if (*param > maxParam) {
+			warnx(
+			    "Invalid warning flag parameter \"%s=%" PRIu32 "\"; capping at maximum %" PRIu8,
+			    rootFlag.c_str(),
+			    *param,
+			    maxParam
+			);
+			*param = maxParam;
+		}
+
+		// Set the first <param> to enabled/error, and disable the rest
+		for (uint32_t ofs = 0; ofs < maxParam; ++ofs) {
+			if (WarningState &warning = state.flagStates[baseID + ofs]; ofs < *param) {
+				warning.update(flagState);
+			} else {
+				warning.state = WARNING_DISABLED;
+			}
+		}
+		return rootFlag;
+	}
+
+	if (param.has_value()) {
+		warnx("Unknown warning flag parameter \"%s=%" PRIu32 "\"", rootFlag.c_str(), *param);
+		return rootFlag;
+	}
+
+	// Try to match against a "meta" warning
+	for (WarningFlag<L> const &metaWarning : metaWarnings) {
+		if (rootFlag != metaWarning.name) {
+			continue;
+		}
+
+		// Set each of the warning flags that meets this level
+		for (W id : EnumSeq(W::NB_WARNINGS)) {
+			if (metaWarning.level >= warningFlags[id].level) {
+				state.metaStates[id].update(flagState);
+			}
+		}
+		return rootFlag;
+	}
+
+	// Try to match against a "normal" flag
+	for (W id : EnumSeq(W::NB_PLAIN_WARNINGS)) {
+		if (rootFlag == warningFlags[id].name) {
+			state.flagStates[id].update(flagState);
+			return rootFlag;
+		}
+	}
+
+	warnx("Unknown warning flag \"%s\"", rootFlag.c_str());
+	return rootFlag;
+}
+
+#endif // RGBDS_DIAGNOSTICS_HPP

include/either.hpp

@@ -1,176 +0,0 @@
-// SPDX-License-Identifier: MIT
-
-#ifndef RGBDS_EITHER_HPP
-#define RGBDS_EITHER_HPP
-
-#include <type_traits>
-#include <utility>
-
-#include "helpers.hpp" // assume
-
-template<typename T1, typename T2>
-union Either {
-	typedef T1 type1;
-	typedef T2 type2;
-
-private:
-	template<typename T, unsigned V>
-	struct Field {
-		constexpr static unsigned tag_value = V;
-
-		unsigned tag = tag_value;
-		T value;
-
-		Field() : value() {}
-		Field(T &value_) : value(value_) {}
-		Field(T const &value_) : value(value_) {}
-		Field(T &&value_) : value(std::move(value_)) {}
-	};
-
-	// The `_tag` unifies with the first `tag` member of each `struct`.
-	constexpr static unsigned nulltag = 0;
-	unsigned _tag = nulltag;
-	Field<T1, 1> _t1;
-	Field<T2, 2> _t2;
-
-	// Value accessors; the function parameters are dummies for overload resolution.
-	// Only used to implement `field()` below.
-	auto &pick(T1 *) { return _t1; }
-	auto const &pick(T1 *) const { return _t1; }
-	auto &pick(T2 *) { return _t2; }
-	auto const &pick(T2 *) const { return _t2; }
-
-	// Generic field accessors; for internal use only.
-	template<typename T>
-	auto &field() {
-		return pick(static_cast<T *>(nullptr));
-	}
-	template<typename T>
-	auto const &field() const {
-		return pick(static_cast<T *>(nullptr));
-	}
-
-public:
-	// Equivalent of `std::monostate` for `std::variant`s.
-	Either() : _tag() {}
-	// These constructors cannot be generic over the value type, because that would prevent
-	// constructible values from being inferred, e.g. a `const char *` string literal for an
-	// `std::string` field value.
-	Either(T1 &value) : _t1(value) {}
-	Either(T2 &value) : _t2(value) {}
-	Either(T1 const &value) : _t1(value) {}
-	Either(T2 const &value) : _t2(value) {}
-	Either(T1 &&value) : _t1(std::move(value)) {}
-	Either(T2 &&value) : _t2(std::move(value)) {}
-
-	// Destructor manually calls the appropriate value destructor.
-	~Either() {
-		if (_tag == _t1.tag_value) {
-			_t1.value.~T1();
-		} else if (_tag == _t2.tag_value) {
-			_t2.value.~T2();
-		}
-	}
-
-	// Copy assignment operators for each possible value.
-	Either &operator=(T1 const &value) {
-		_t1.tag = _t1.tag_value;
-		new (&_t1.value) T1(value);
-		return *this;
-	}
-	Either &operator=(T2 const &value) {
-		_t2.tag = _t2.tag_value;
-		new (&_t2.value) T2(value);
-		return *this;
-	}
-
-	// Move assignment operators for each possible value.
-	Either &operator=(T1 &&value) {
-		_t1.tag = _t1.tag_value;
-		new (&_t1.value) T1(std::move(value));
-		return *this;
-	}
-	Either &operator=(T2 &&value) {
-		_t2.tag = _t2.tag_value;
-		new (&_t2.value) T2(std::move(value));
-		return *this;
-	}
-
-	// Copy assignment operator from another `Either`.
-	Either &operator=(Either other) {
-		if (other._tag == other._t1.tag_value) {
-			*this = other._t1.value;
-		} else if (other._tag == other._t2.tag_value) {
-			*this = other._t2.value;
-		} else {
-			_tag = nulltag; // LCOV_EXCL_LINE
-		}
-		return *this;
-	}
-
-	// Copy constructor from another `Either`; implemented in terms of value assignment operators.
-	Either(Either const &other) {
-		if (other._tag == other._t1.tag_value) {
-			*this = other._t1.value;
-		} else if (other._tag == other._t2.tag_value) {
-			*this = other._t2.value;
-		} else {
-			_tag = nulltag;
-		}
-	}
-
-	// Move constructor from another `Either`; implemented in terms of value assignment operators.
-	Either(Either &&other) {
-		if (other._tag == other._t1.tag_value) {
-			*this = std::move(other._t1.value);
-		} else if (other._tag == other._t2.tag_value) {
-			*this = std::move(other._t2.value);
-		} else {
-			_tag = nulltag;
-		}
-	}
-
-	// Equivalent of `.emplace<T>()` for `std::variant`s.
-	template<typename T, typename... Args>
-	void emplace(Args &&...args) {
-		this->~Either();
-		if constexpr (std::is_same_v<T, T1>) {
-			_t1.tag = _t1.tag_value;
-			new (&_t1.value) T1(std::forward<Args>(args)...);
-		} else if constexpr (std::is_same_v<T, T2>) {
-			_t2.tag = _t2.tag_value;
-			new (&_t2.value) T2(std::forward<Args>(args)...);
-		} else {
-			_tag = nulltag;
-		}
-	}
-
-	// Equivalent of `std::holds_alternative<std::monostate>()` for `std::variant`s.
-	bool empty() const { return _tag == nulltag; }
-
-	// Equivalent of `std::holds_alternative<T>()` for `std::variant`s.
-	template<typename T>
-	bool holds() const {
-		if constexpr (std::is_same_v<T, T1>) {
-			return _tag == _t1.tag_value;
-		} else if constexpr (std::is_same_v<T, T2>) {
-			return _tag == _t2.tag_value;
-		} else {
-			return false;
-		}
-	}
-
-	// Equivalent of `std::get<T>()` for `std::variant`s.
-	template<typename T>
-	auto &get() {
-		assume(holds<T>());
-		return field<T>().value;
-	}
-	template<typename T>
-	auto const &get() const {
-		assume(holds<T>());
-		return field<T>().value;
-	}
-};
-
-#endif // RGBDS_EITHER_HPP

include/error.hpp

@@ -1,18 +0,0 @@
-// SPDX-License-Identifier: MIT
-
-#ifndef RGBDS_ERROR_HPP
-#define RGBDS_ERROR_HPP
-
-extern "C" {
-	[[gnu::format(printf, 1, 2)]]
-	void warn(char const *fmt...);
-	[[gnu::format(printf, 1, 2)]]
-	void warnx(char const *fmt, ...);
-
-	[[gnu::format(printf, 1, 2), noreturn]]
-	void err(char const *fmt, ...);
-	[[gnu::format(printf, 1, 2), noreturn]]
-	void errx(char const *fmt, ...);
-}
-
-#endif // RGBDS_ERROR_HPP

include/extern/utf8decoder.hpp

@@ -5,6 +5,9 @@
 
 #include <stdint.h>
 
+#define UTF8_ACCEPT 0
+#define UTF8_REJECT 12
+
 uint32_t decode(uint32_t *state, uint32_t *codep, uint8_t byte);
 
 #endif // RGBDS_EXTERN_UTF8DECODER_HPP

include/file.hpp

@@ -10,15 +10,13 @@
 #include <streambuf>
 #include <string.h>
 #include <string>
+#include <variant>
 
-#include "either.hpp"
 #include "helpers.hpp" // assume
 #include "platform.hpp"
 
-#include "gfx/main.hpp"
-
 class File {
-	Either<std::streambuf *, std::filebuf> _file;
+	std::variant<std::streambuf *, std::filebuf> _file;
 
 public:
 	File() : _file(nullptr) {}
@@ -27,17 +25,12 @@ public:
 	// Returns `nullptr` on error, and a non-null pointer otherwise.
 	File *open(std::string const &path, std::ios_base::openmode mode) {
 		if (path != "-") {
-			_file.emplace<std::filebuf>();
-			return _file.get<std::filebuf>().open(path, mode) ? this : nullptr;
+			return _file.emplace<std::filebuf>().open(path, mode) ? this : nullptr;
 		} else if (mode & std::ios_base::in) {
 			assume(!(mode & std::ios_base::out));
 			_file.emplace<std::streambuf *>(std::cin.rdbuf());
 			if (setmode(STDIN_FILENO, (mode & std::ios_base::binary) ? O_BINARY : O_TEXT) == -1) {
-				fatal(
-				    "Failed to set stdin to %s mode: %s",
-				    mode & std::ios_base::binary ? "binary" : "text",
-				    strerror(errno)
-				);
+				return nullptr;
 			}
 		} else {
 			assume(mode & std::ios_base::out);
@@ -46,8 +39,8 @@ public:
 		return this;
 	}
 	std::streambuf &operator*() {
-		return _file.holds<std::filebuf>() ? _file.get<std::filebuf>()
-		                                   : *_file.get<std::streambuf *>();
+		return std::holds_alternative<std::filebuf>(_file) ? std::get<std::filebuf>(_file)
+		                                                   : *std::get<std::streambuf *>(_file);
 	}
 	std::streambuf const &operator*() const {
 		// The non-`const` version does not perform any modifications, so it's okay.
@@ -60,8 +53,8 @@ public:
 	}
 
 	char const *c_str(std::string const &path) const {
-		return _file.holds<std::filebuf>()                         ? path.c_str()
-		       : _file.get<std::streambuf *>() == std::cin.rdbuf() ? "<stdin>"
+		return std::holds_alternative<std::filebuf>(_file)             ? path.c_str()
+		       : std::get<std::streambuf *>(_file) == std::cin.rdbuf() ? "<stdin>"
 		                                                               : "<stdout>";
 	}
 };

include/fix/mbc.hpp

@@ -0,0 +1,79 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_FIX_MBC_HPP
+#define RGBDS_FIX_MBC_HPP
+
+#include <stdint.h>
+#include <stdio.h>
+
+constexpr uint16_t UNSPECIFIED = 0x200;
+static_assert(UNSPECIFIED > 0xFF, "UNSPECIFIED should not be in byte range!");
+
+enum MbcType {
+	ROM = 0x00,
+	ROM_RAM = 0x08,
+	ROM_RAM_BATTERY = 0x09,
+
+	MBC1 = 0x01,
+	MBC1_RAM = 0x02,
+	MBC1_RAM_BATTERY = 0x03,
+
+	MBC2 = 0x05,
+	MBC2_BATTERY = 0x06,
+
+	MMM01 = 0x0B,
+	MMM01_RAM = 0x0C,
+	MMM01_RAM_BATTERY = 0x0D,
+
+	MBC3 = 0x11,
+	MBC3_TIMER_BATTERY = 0x0F,
+	MBC3_TIMER_RAM_BATTERY = 0x10,
+	MBC3_RAM = 0x12,
+	MBC3_RAM_BATTERY = 0x13,
+
+	MBC5 = 0x19,
+	MBC5_RAM = 0x1A,
+	MBC5_RAM_BATTERY = 0x1B,
+	MBC5_RUMBLE = 0x1C,
+	MBC5_RUMBLE_RAM = 0x1D,
+	MBC5_RUMBLE_RAM_BATTERY = 0x1E,
+
+	MBC6 = 0x20,
+
+	MBC7_SENSOR_RUMBLE_RAM_BATTERY = 0x22,
+
+	POCKET_CAMERA = 0xFC,
+
+	BANDAI_TAMA5 = 0xFD,
+
+	HUC3 = 0xFE,
+
+	HUC1_RAM_BATTERY = 0xFF,
+
+	// "Extended" values (still valid, but not directly actionable)
+
+	// A high byte of 0x01 means TPP1, the low byte is the requested features
+	// This does not include SRAM, which is instead implied by a non-zero SRAM size
+	// Note: Multiple rumble speeds imply rumble
+	TPP1 = 0x100,
+	TPP1_RUMBLE = 0x101,
+	TPP1_MULTIRUMBLE_RUMBLE = 0x103,
+	TPP1_TIMER = 0x104,
+	TPP1_TIMER_RUMBLE = 0x105,
+	TPP1_TIMER_MULTIRUMBLE_RUMBLE = 0x107,
+	TPP1_BATTERY = 0x108,
+	TPP1_BATTERY_RUMBLE = 0x109,
+	TPP1_BATTERY_MULTIRUMBLE_RUMBLE = 0x10B,
+	TPP1_BATTERY_TIMER = 0x10C,
+	TPP1_BATTERY_TIMER_RUMBLE = 0x10D,
+	TPP1_BATTERY_TIMER_MULTIRUMBLE_RUMBLE = 0x10F,
+
+	// Error values
+	MBC_NONE = UNSPECIFIED, // No MBC specified, do not act on it
+};
+
+bool mbc_HasRAM(MbcType type);
+char const *mbc_Name(MbcType type);
+MbcType mbc_ParseName(char const *name, uint8_t &tpp1Major, uint8_t &tpp1Minor);
+
+#endif // RGBDS_FIX_MBC_HPP

include/fix/warning.hpp

@@ -0,0 +1,41 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_FIX_WARNING_HPP
+#define RGBDS_FIX_WARNING_HPP
+
+#include "diagnostics.hpp"
+
+enum WarningLevel {
+	LEVEL_DEFAULT,    // Warnings that are enabled by default
+	LEVEL_ALL,        // Warnings that probably indicate an error
+	LEVEL_EVERYTHING, // Literally every warning
+};
+
+enum WarningID {
+	WARNING_MBC,        // Issues with MBC specs
+	WARNING_OVERWRITE,  // Overwriting non-zero bytes
+	WARNING_SGB,        // SGB flag conflicts with old licensee code
+	WARNING_TRUNCATION, // Truncating values to fit
+
+	NB_PLAIN_WARNINGS,
+
+	NB_WARNINGS = NB_PLAIN_WARNINGS,
+};
+
+extern Diagnostics<WarningLevel, WarningID> warnings;
+
+// Warns the user about problems that don't prevent fixing the ROM header
+[[gnu::format(printf, 2, 3)]]
+void warning(WarningID id, char const *fmt, ...);
+
+// Prints an error, and increments the error count
+[[gnu::format(printf, 1, 2)]]
+void error(char const *fmt, ...);
+
+// Prints an error, and exits with failure
+[[gnu::format(printf, 1, 2), noreturn]]
+void fatal(char const *fmt, ...);
+
+uint32_t checkErrors(char const *filename);
+
+#endif // RGBDS_FIX_WARNING_HPP

include/gfx/color_set.hpp

@@ -1,13 +1,13 @@
 // SPDX-License-Identifier: MIT
 
-#ifndef RGBDS_GFX_PROTO_PALETTE_HPP
-#define RGBDS_GFX_PROTO_PALETTE_HPP
+#ifndef RGBDS_GFX_COLOR_SET_HPP
+#define RGBDS_GFX_COLOR_SET_HPP
 
 #include <array>
 #include <stddef.h>
 #include <stdint.h>
 
-class ProtoPalette {
+class ColorSet {
 public:
 	static constexpr size_t capacity = 4;
 
@@ -26,7 +26,7 @@ public:
 		WE_BIGGER,
 		THEY_BIGGER = -1,
 	};
-	ComparisonResult compare(ProtoPalette const &other) const;
+	ComparisonResult compare(ColorSet const &other) const;
 
 	size_t size() const;
 	bool empty() const;
@@ -35,4 +35,4 @@ public:
 	decltype(_colorIndices)::const_iterator end() const;
 };
 
-#endif // RGBDS_GFX_PROTO_PALETTE_HPP
+#endif // RGBDS_GFX_COLOR_SET_HPP

include/gfx/main.hpp

@@ -40,7 +40,10 @@ struct Options {
 		uint16_t top;
 		uint16_t width;
 		uint16_t height;
+		uint32_t right() const { return left + width * 8; }
+		uint32_t bottom() const { return top + height * 8; }
 	} inputSlice{0, 0, 0, 0};                          // -L (margins in clockwise order, like CSS)
+	uint8_t basePalID = 0;                             // -l
 	std::array<uint16_t, 2> maxNbTiles{UINT16_MAX, 0}; // -N
 	uint16_t nbPalettes = 8;                           // -n
 	std::string output{};                              // -o
@@ -68,6 +71,8 @@ struct Options {
 	mutable bool hasTransparentPixels = false;
 	uint8_t maxOpaqueColors() const { return nbColorsPerPal - hasTransparentPixels; }
 
+	uint16_t maxNbColors() const { return nbColorsPerPal * nbPalettes; }
+
 	uint8_t dmgColors[4] = {};
 	uint8_t dmgValue(uint8_t i) const {
 		assume(i < 4);
@@ -77,25 +82,6 @@ struct Options {
 
 extern Options options;
 
-// Prints the error count, and exits with failure
-[[noreturn]]
-void giveUp();
-// If any error has been emitted thus far, calls `giveUp()`.
-void requireZeroErrors();
-// Prints a warning, and does not change the error count
-[[gnu::format(printf, 1, 2)]]
-void warning(char const *fmt, ...);
-// Prints an error, and increments the error count
-[[gnu::format(printf, 1, 2)]]
-void error(char const *fmt, ...);
-// Prints an error, and increments the error count
-// Does not take format arguments so `format_` and `-Wformat-security` won't complain about
-// calling `errorMessage(msg)`.
-void errorMessage(char const *msg);
-// Prints a fatal error, increments the error count, and gives up
-[[gnu::format(printf, 1, 2), noreturn]]
-void fatal(char const *fmt, ...);
-
 struct Palette {
 	// An array of 4 GBC-native (RGB555) colors
 	std::array<uint16_t, 4> colors{UINT16_MAX, UINT16_MAX, UINT16_MAX, UINT16_MAX};
@@ -114,9 +100,9 @@ struct Palette {
 };
 
 // Flipping tends to happen fairly often, so take a bite out of dcache to speed it up
-static constexpr auto flipTable = ([]() constexpr {
+static std::array<uint16_t, 256> flipTable = ([]() constexpr {
 	std::array<uint16_t, 256> table{};
-	for (uint16_t i = 0; i < table.size(); i++) {
+	for (uint16_t i = 0; i < table.size(); ++i) {
 		// To flip all the bits, we'll flip both nibbles, then each nibble half, etc.
 		uint16_t byte = i;
 		byte = (byte & 0b0000'1111) << 4 | (byte & 0b1111'0000) >> 4;

include/gfx/pal_packing.hpp

@@ -3,16 +3,14 @@
 #ifndef RGBDS_GFX_PAL_PACKING_HPP
 #define RGBDS_GFX_PAL_PACKING_HPP
 
+#include <stddef.h>
 #include <tuple>
 #include <vector>
 
-#include "defaultinitvec.hpp"
-
 struct Palette;
-class ProtoPalette;
+class ColorSet;
 
-// Returns which palette each proto-palette maps to, and how many palettes are necessary
-std::tuple<DefaultInitVec<size_t>, size_t>
-    overloadAndRemove(std::vector<ProtoPalette> const &protoPalettes);
+// Returns which palette each color set maps to, and how many palettes are necessary
+std::tuple<std::vector<size_t>, size_t> overloadAndRemove(std::vector<ColorSet> const &colorSets);
 
 #endif // RGBDS_GFX_PAL_PACKING_HPP

include/gfx/pal_sorting.hpp

@@ -5,7 +5,6 @@
 
 #include <array>
 #include <optional>
-#include <png.h>
 #include <vector>
 
 #include "gfx/rgba.hpp"
@@ -16,13 +15,7 @@ static constexpr size_t NB_COLOR_SLOTS = (1 << (5 * 3)) + 1;
 
 struct Palette;
 
-void sortIndexed(
-    std::vector<Palette> &palettes,
-    int palSize,
-    png_color const *palRGB,
-    int palAlphaSize,
-    png_byte *palAlpha
-);
+void sortIndexed(std::vector<Palette> &palettes, std::vector<Rgba> const &embPal);
 void sortGrayscale(
     std::vector<Palette> &palettes, std::array<std::optional<Rgba>, NB_COLOR_SLOTS> const &colors
 );

include/gfx/png.hpp

@@ -0,0 +1,21 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_GFX_PNG_HPP
+#define RGBDS_GFX_PNG_HPP
+
+#include <fstream>
+#include <stdint.h>
+#include <vector>
+
+#include "gfx/rgba.hpp"
+
+struct Png {
+	uint32_t width, height;
+	std::vector<Rgba> pixels{};
+	std::vector<Rgba> palette{};
+
+	Png() {}
+	Png(char const *filename, std::streambuf &file);
+};
+
+#endif // RGBDS_GFX_PNG_HPP

include/gfx/rgba.hpp

@@ -17,23 +17,25 @@ struct Rgba {
 	explicit constexpr Rgba(uint32_t rgba = 0)
 	    : red(rgba >> 24), green(rgba >> 16), blue(rgba >> 8), alpha(rgba) {}
 
-	static constexpr Rgba fromCGBColor(uint16_t cgbColor) {
-		constexpr auto _5to8 = [](uint8_t fiveBpp) -> uint8_t {
-			fiveBpp &= 0b11111; // For caller's convenience
-			return fiveBpp << 3 | fiveBpp >> 2;
+	static constexpr Rgba fromCGBColor(uint16_t color) {
+		constexpr auto _5to8 = [](uint8_t channel) -> uint8_t {
+			channel &= 0b11111; // For caller's convenience
+			return channel << 3 | channel >> 2;
 		};
 		return {
-		    _5to8(cgbColor),
-		    _5to8(cgbColor >> 5),
-		    _5to8(cgbColor >> 10),
-		    static_cast<uint8_t>(cgbColor & 0x8000 ? 0x00 : 0xFF),
+		    _5to8(color),
+		    _5to8(color >> 5),
+		    _5to8(color >> 10),
+		    static_cast<uint8_t>(color & 0x8000 ? 0x00 : 0xFF),
 		};
 	}
 
 	// Returns this RGBA as a 32-bit number that can be printed in hex (`%08x`) to yield its CSS
 	// representation
 	uint32_t toCSS() const {
-		auto shl = [](uint8_t val, unsigned shift) { return static_cast<uint32_t>(val) << shift; };
+		constexpr auto shl = [](uint8_t val, unsigned shift) {
+			return static_cast<uint32_t>(val) << shift;
+		};
 		return shl(red, 24) | shl(green, 16) | shl(blue, 8) | shl(alpha, 0);
 	}
 

include/gfx/warning.hpp

@@ -0,0 +1,44 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_GFX_WARNING_HPP
+#define RGBDS_GFX_WARNING_HPP
+
+#include "diagnostics.hpp"
+
+enum WarningLevel {
+	LEVEL_DEFAULT,    // Warnings that are enabled by default
+	LEVEL_ALL,        // Warnings that probably indicate an error
+	LEVEL_EVERYTHING, // Literally every warning
+};
+
+enum WarningID {
+	WARNING_EMBEDDED,      // Using an embedded PNG palette without '-c embedded'
+	WARNING_TRIM_NONEMPTY, // '-x' trims nonempty tiles
+
+	NB_PLAIN_WARNINGS,
+
+	NB_WARNINGS = NB_PLAIN_WARNINGS,
+};
+
+extern Diagnostics<WarningLevel, WarningID> warnings;
+
+// Warns the user about problems that don't prevent valid graphics conversion
+[[gnu::format(printf, 2, 3)]]
+void warning(WarningID id, char const *fmt, ...);
+
+// Prints the error count, and exits with failure
+[[noreturn]]
+void giveUp();
+
+// If any error has been emitted thus far, calls `giveUp()`
+void requireZeroErrors();
+
+// Prints an error, and increments the error count
+[[gnu::format(printf, 1, 2)]]
+void error(char const *fmt, ...);
+
+// Prints a fatal error, increments the error count, and gives up
+[[gnu::format(printf, 1, 2), noreturn]]
+void fatal(char const *fmt, ...);
+
+#endif // RGBDS_GFX_WARNING_HPP

include/helpers.hpp

@@ -71,7 +71,7 @@ static inline int ctz(unsigned int x) {
 
 	while (!(x & 1)) {
 		x >>= 1;
-		cnt++;
+		++cnt;
 	}
 	return cnt;
 }
@@ -81,7 +81,7 @@ static inline int clz(unsigned int x) {
 
 	while (x <= UINT_MAX / 2) {
 		x <<= 1;
-		cnt++;
+		++cnt;
 	}
 	return cnt;
 }

include/itertools.hpp

@@ -22,7 +22,7 @@ class EnumSeq {
 			return *this;
 		}
 
-		auto operator*() const { return _value; }
+		T operator*() const { return _value; }
 
 		bool operator==(Iterator const &rhs) const { return _value == rhs._value; }
 	};

include/link/assign.hpp

@@ -3,10 +3,6 @@
 #ifndef RGBDS_LINK_ASSIGN_HPP
 #define RGBDS_LINK_ASSIGN_HPP
 
-#include <stdint.h>
-
-extern uint64_t nbSectionsToAssign;
-
 // Assigns all sections a slice of the address space
 void assign_AssignSections();
 

include/link/layout.hpp

@@ -0,0 +1,22 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_LINK_LAYOUT_HPP
+#define RGBDS_LINK_LAYOUT_HPP
+
+#include <stdint.h>
+#include <string>
+
+#include "linkdefs.hpp"
+
+void layout_SetFloatingSectionType(SectionType type);
+void layout_SetSectionType(SectionType type);
+void layout_SetSectionType(SectionType type, uint32_t bank);
+
+void layout_SetAddr(uint32_t addr);
+void layout_MakeAddrFloating();
+void layout_AlignTo(uint32_t alignment, uint32_t offset);
+void layout_Pad(uint32_t length);
+
+void layout_PlaceSection(std::string const &name, bool isOptional);
+
+#endif // RGBDS_LINK_LAYOUT_HPP

include/link/lexer.hpp

@@ -0,0 +1,17 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_LINK_LEXER_HPP
+#define RGBDS_LINK_LEXER_HPP
+
+#include <stdarg.h>
+#include <string>
+
+[[gnu::format(printf, 1, 2)]]
+void lexer_Error(char const *fmt, ...);
+
+void lexer_IncludeFile(std::string &&path);
+void lexer_IncLineNo();
+
+bool lexer_Init(char const *linkerScriptName);
+
+#endif // RGBDS_LINK_LEXER_HPP

include/link/main.hpp

@@ -6,40 +6,43 @@
 #include <stdint.h>
 #include <stdio.h>
 #include <string>
+#include <variant>
 #include <vector>
 
-#include "either.hpp"
 #include "linkdefs.hpp"
 
-// Variables related to CLI options
-extern bool isDmgMode;
-extern char const *linkerScriptName;
-extern char const *mapFileName;
-extern bool noSymInMap;
-extern char const *symFileName;
-extern char const *overlayFileName;
-extern char const *outputFileName;
-extern uint8_t padValue;
-extern bool hasPadValue;
-extern uint16_t scrambleROMX;
-extern uint8_t scrambleWRAMX;
-extern uint8_t scrambleSRAM;
-extern bool is32kMode;
-extern bool beVerbose;
-extern bool isWRAM0Mode;
-extern bool disablePadding;
+struct Options {
+	bool isDmgMode;              // -d
+	char const *mapFileName;     // -m
+	bool noSymInMap;             // -M
+	char const *symFileName;     // -n
+	char const *overlayFileName; // -O
+	char const *outputFileName;  // -o
+	uint8_t padValue;            // -p
+	bool hasPadValue = false;
+	// Setting these three to 0 disables the functionality
+	uint16_t scrambleROMX; // -S
+	uint16_t scrambleWRAMX;
+	uint16_t scrambleSRAM;
+	bool is32kMode;      // -t
+	bool beVerbose;      // -v
+	bool isWRAM0Mode;    // -w
+	bool disablePadding; // -x
+};
+
+extern Options options;
 
-// Helper macro for printing verbose-mode messages
 #define verbosePrint(...) \
 	do { \
-		if (beVerbose) { \
+		if (options.beVerbose) { \
 			fprintf(stderr, __VA_ARGS__); \
 		} \
 	} while (0)
 
 struct FileStackNode {
 	FileStackNodeType type;
-	Either<
+	std::variant<
+	    std::monostate,        // Default constructed; `.type` and `.data` must be set manually
 	    std::vector<uint32_t>, // NODE_REPT
 	    std::string            // NODE_FILE, NODE_MACRO
 	    >
@@ -50,20 +53,13 @@ struct FileStackNode {
 	uint32_t lineNo;
 
 	// REPT iteration counts since last named node, in reverse depth order
-	std::vector<uint32_t> &iters() { return data.get<std::vector<uint32_t>>(); }
-	std::vector<uint32_t> const &iters() const { return data.get<std::vector<uint32_t>>(); }
+	std::vector<uint32_t> &iters() { return std::get<std::vector<uint32_t>>(data); }
+	std::vector<uint32_t> const &iters() const { return std::get<std::vector<uint32_t>>(data); }
 	// File name for files, file::macro name for macros
-	std::string &name() { return data.get<std::string>(); }
-	std::string const &name() const { return data.get<std::string>(); }
+	std::string &name() { return std::get<std::string>(data); }
+	std::string const &name() const { return std::get<std::string>(data); }
 
 	std::string const &dump(uint32_t curLineNo) const;
 };
 
-[[gnu::format(printf, 3, 4)]]
-void warning(FileStackNode const *where, uint32_t lineNo, char const *fmt, ...);
-[[gnu::format(printf, 3, 4)]]
-void error(FileStackNode const *where, uint32_t lineNo, char const *fmt, ...);
-[[gnu::format(printf, 3, 4), noreturn]]
-void fatal(FileStackNode const *where, uint32_t lineNo, char const *fmt, ...);
-
 #endif // RGBDS_LINK_MAIN_HPP

include/link/patch.hpp

@@ -3,6 +3,23 @@
 #ifndef RGBDS_LINK_PATCH_HPP
 #define RGBDS_LINK_PATCH_HPP
 
+#include <string>
+#include <vector>
+
+#include "link/section.hpp"
+
+struct Symbol;
+
+struct Assertion {
+	Patch patch; // Also used for its `.type`
+	std::string message;
+	// This would be redundant with `patch.pcSection->fileSymbols`, but `section` is sometimes
+	// `nullptr`!
+	std::vector<Symbol> *fileSymbols;
+};
+
+Assertion &patch_AddAssertion();
+
 // Checks all assertions
 void patch_CheckAssertions();
 

include/link/section.hpp

@@ -56,15 +56,6 @@ struct Section {
 	std::unique_ptr<Section> nextu; // The next "component" of this unionized sect
 };
 
-struct Assertion {
-	Patch patch; // Also used for its `.type`
-	std::string message;
-	// This would be redundant with `.section->fileSymbols`, but `section` is sometimes `nullptr`!
-	std::vector<Symbol> *fileSymbols;
-};
-
-extern std::deque<Assertion> assertions;
-
 // Execute a callback for each section currently registered.
 // This is to avoid exposing the data structure in which sections are stored.
 void sect_ForEach(void (*callback)(Section &));

include/link/symbol.hpp

@@ -7,8 +7,8 @@
 
 #include <stdint.h>
 #include <string>
+#include <variant>
 
-#include "either.hpp"
 #include "linkdefs.hpp"
 
 struct FileStackNode;
@@ -27,14 +27,14 @@ struct Symbol {
 	ExportLevel type;
 	FileStackNode const *src;
 	int32_t lineNo;
-	Either<
+	std::variant<
 	    int32_t, // Constants just have a numeric value
 	    Label    // Label values refer to an offset within a specific section
 	    >
 	    data;
 
-	Label &label() { return data.get<Label>(); }
-	Label const &label() const { return data.get<Label>(); }
+	Label &label() { return std::get<Label>(data); }
+	Label const &label() const { return std::get<Label>(data); }
 };
 
 void sym_ForEach(void (*callback)(Symbol &));

include/link/warning.hpp

@@ -0,0 +1,61 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_LINK_WARNING_HPP
+#define RGBDS_LINK_WARNING_HPP
+
+#include <stdarg.h>
+#include <stdint.h>
+
+#include "diagnostics.hpp"
+
+#define warningAt(where, ...) warning(where.src, where.lineNo, __VA_ARGS__)
+#define errorAt(where, ...)   error(where.src, where.lineNo, __VA_ARGS__)
+#define fatalAt(where, ...)   fatal(where.src, where.lineNo, __VA_ARGS__)
+
+enum WarningLevel {
+	LEVEL_DEFAULT,    // Warnings that are enabled by default
+	LEVEL_ALL,        // Warnings that probably indicate an error
+	LEVEL_EVERYTHING, // Literally every warning
+};
+
+enum WarningID {
+	WARNING_ASSERT,       // Assertions
+	WARNING_DIV,          // Undefined division behavior
+	WARNING_OBSOLETE,     // Obsolete/deprecated things
+	WARNING_SHIFT,        // Undefined `SHIFT` behavior
+	WARNING_SHIFT_AMOUNT, // Strange `SHIFT` amount
+	WARNING_TRUNCATION,   // Implicit truncation loses some bits
+
+	NB_PLAIN_WARNINGS,
+
+	NB_WARNINGS = NB_PLAIN_WARNINGS,
+};
+
+extern Diagnostics<WarningLevel, WarningID> warnings;
+
+struct FileStackNode;
+
+[[gnu::format(printf, 4, 5)]]
+void warning(FileStackNode const *src, uint32_t lineNo, WarningID id, char const *fmt, ...);
+[[gnu::format(printf, 3, 4)]]
+void warning(FileStackNode const *src, uint32_t lineNo, char const *fmt, ...);
+[[gnu::format(printf, 1, 2)]]
+void warning(char const *fmt, ...);
+
+[[gnu::format(printf, 3, 4)]]
+void error(FileStackNode const *src, uint32_t lineNo, char const *fmt, ...);
+[[gnu::format(printf, 1, 2)]]
+void error(char const *fmt, ...);
+[[gnu::format(printf, 1, 2)]]
+void errorNoDump(char const *fmt, ...);
+
+void scriptError(char const *name, uint32_t lineNo, char const *fmt, va_list args);
+
+[[gnu::format(printf, 3, 4), noreturn]]
+void fatal(FileStackNode const *src, uint32_t lineNo, char const *fmt, ...);
+[[gnu::format(printf, 1, 2), noreturn]]
+void fatal(char const *fmt, ...);
+
+void requireZeroErrors();
+
+#endif // RGBDS_LINK_WARNING_HPP

include/platform.hpp

@@ -1,7 +1,5 @@
 // SPDX-License-Identifier: MIT
 
-// platform-specific hacks
-
 #ifndef RGBDS_PLATFORM_HPP
 #define RGBDS_PLATFORM_HPP
 

include/usage.hpp

@@ -0,0 +1,21 @@
+// SPDX-License-Identifier: MIT
+
+#ifndef RGBDS_USAGE_HPP
+#define RGBDS_USAGE_HPP
+
+#include <stdarg.h>
+
+class Usage {
+	char const *usage;
+
+public:
+	Usage(char const *usage_) : usage(usage_) {}
+
+	[[noreturn]]
+	void printAndExit(int code) const;
+
+	[[gnu::format(printf, 2, 3), noreturn]]
+	void printAndExit(char const *fmt, ...) const;
+};
+
+#endif // RGBDS_USAGE_HPP

include/util.hpp

@@ -3,9 +3,23 @@
 #ifndef RGBDS_UTIL_HPP
 #define RGBDS_UTIL_HPP
 
+#include <stddef.h>
+#include <string>
+#include <unordered_map>
+
+#include "helpers.hpp"
+
 bool startsIdentifier(int c);
 bool continuesIdentifier(int c);
 
 char const *printChar(int c);
 
+struct Uppercase {
+	size_t operator()(std::string const &str) const;
+	bool operator()(std::string const &str1, std::string const &str2) const;
+};
+
+template<typename T>
+using UpperMap = std::unordered_map<std::string, T, Uppercase, Uppercase>;
+
 #endif // RGBDS_UTIL_HPP

include/version.hpp

@@ -5,7 +5,7 @@
 
 #define PACKAGE_VERSION_MAJOR 0
 #define PACKAGE_VERSION_MINOR 9
-#define PACKAGE_VERSION_PATCH 3
+#define PACKAGE_VERSION_PATCH 4
 
 char const *get_package_version_string();
 

man/gbz80.7

@@ -1,6 +1,6 @@
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt GBZ80 7
 .Os
 .Sh NAME

man/rgbasm-old.5

@@ -2,7 +2,7 @@
 .\"
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBASM-OLD 5
 .Os
 .Sh NAME
@@ -99,6 +99,23 @@ in RAM sections, allowing labeled space allocations to overlap.
 .Pp
 Instead, use
 .Ic UNION .
+.Ss Section-local charmaps
+Deprecated in 0.3.9, removed in 0.4.0.
+.Pp
+Defining a
+.Ic CHARMAP
+inside a
+.Ic SECTION
+when the current global charmap was the
+.Sq main
+one used to only define that character mapping within that
+.Ic SECTION .
+.Pp
+Instead, use
+.Ic PUSHC
+and
+.Ic POPC
+and switch to a different character mapping for that section.
 .Ss __FILE__ and __LINE__
 Deprecated in 0.6.0, removed in 0.7.0.
 .Pp
@@ -301,6 +318,11 @@ Supported in ASMotor, removed in RGBDS.
 .Pp
 Instead, use
 .Ql LD HL, SP + e8 .
+.Ss OPT z
+Deprecated in 0.4.0, removed in 0.5.0.
+.Pp
+Instead, use
+.Ic OPT p .
 .Ss rgbasm -i
 Deprecated in 0.6.0, removed in 0.8.0.
 .Pp
@@ -356,6 +378,36 @@ because $16384$ turns = $16384 tau$ radians = $32768 pi$ radians, and $sin ( 327
 .EQ
 delim off
 .EN
+.Ss % operator behavior with negative dividend or divisor
+Changed in 0.5.0.
+.Pp
+Instead of having the same sign as the dividend (a remainder operation),
+.Ql %
+has the same sign as the divisor (a modulo operation).
+.Pp
+For example, previously we had:
+.Bl -bullet -offset indent
+.It
+.Ql 13 % 10 == 3
+.It
+.Ql -13 % 10 == -3
+.It
+.Ql 13 % -10 == 3
+.It
+.Ql -13 % -10 == -3
+.El
+.Pp
+Instead, now we have:
+.Bl -bullet -offset indent
+.It
+.Ql 13 % 10 == 3
+.It
+.Ql -13 % 10 == 7
+.It
+.Ql 13 % -10 == -7
+.It
+.Ql -13 % -10 == -3
+.El
 .Ss ** operator associativity
 Changed in 0.9.0.
 .Pp

man/rgbasm.1

@@ -1,6 +1,6 @@
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBASM 1
 .Os
 .Sh NAME
@@ -15,6 +15,7 @@
 .Op Fl I Ar path
 .Op Fl M Ar depend_file
 .Op Fl MG
+.Op Fl MC
 .Op Fl MP
 .Op Fl MT Ar target_file
 .Op Fl MQ Ar target_file
@@ -99,11 +100,12 @@ Add a new
 .Dq include path ;
 .Ar path
 must point to a directory.
-When a
+When any
 .Ic INCLUDE
-.Pq including the implicit one from Fl P
+.Pq including the implicit one from Fl P ,
+.Ic INCBIN ,
 or
-.Ic INCBIN
+.Ic READFILE
 is attempted,
 .Nm
 first looks up the provided path from its working directory; if this fails, it tries again from each of the
@@ -119,17 +121,33 @@ To be used in conjunction with
 .Fl M .
 This makes
 .Nm
-assume that missing files are auto-generated: when
+assume that missing files are auto-generated: when any
 .Ic INCLUDE
-.Pq including the implicit one from Fl P
+.Pq including the implicit one from Fl P ,
+.Ic INCBIN ,
 or
-.Ic INCBIN
+.Ic READFILE
 is attempted on a non-existent file, it is added as a dependency, then
 .Nm
-exits normally instead of erroring out.
-This feature is used in automatic updating of makefiles.
+exits normally or continues processing (depending on whether
+.Fl MC
+was enabled) instead of erroring out.
+This feature is used in automatic updating of Makefiles.
+.It Fl MC
+Implies
+.Fl MG .
+This makes
+.Nm
+continue processing after a non-existent dependency file, instead of exiting.
+Note that this is
+.Em not
+recommended if any non-existent dependencies would have influenced subsequent processing, e.g. by causing an
+.Ic IF
+condition to take a different branch.
 .It Fl MP
-When enabled, this causes a phony target to be added for each dependency other than the main file.
+When enabled, this adds a phony target to the rules emitted by
+.Fl M
+for each dependency other than the main file.
 This prevents
 .Xr make 1
 from erroring out when dependency files are deleted.
@@ -158,6 +176,7 @@ This acts as if a
 .Ql Ic INCLUDE Qq Ar include_file
 was read before the input
 .Ar asmfile .
+Multiple files can be pre-included in the order they were provided.
 .It Fl p Ar pad_value , Fl \-pad-value Ar pad_value
 Use this as the value for
 .Ic DS
@@ -269,7 +288,7 @@ Enables literally every warning.
 .El
 .Pp
 The following warnings are actual warning flags; with each description, the corresponding warning flag is included.
-Note that each of these flag also has a negation (for example,
+Note that each of these flags also has a negation (for example,
 .Fl Wcharmap-redef
 enables the warning that
 .Fl Wno-charmap-redef
@@ -336,11 +355,7 @@ Block comments cannot be nested, so the first
 .Ql */
 will end the whole comment.
 .It Fl Wno-obsolete
-Warn when obsolete constructs such as the
-.Ic _PI
-constant or
-.Ic PRINTT
-directive are encountered.
+Warn when obsolete features are encountered, which have been deprecated and may later be removed.
 .It Fl Wnumeric-string=
 Warn when a multi-character string is treated as a number.
 .Fl Wnumeric-string=0

man/rgbasm.5

@@ -2,7 +2,7 @@
 .\"
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBASM 5
 .Os
 .Sh NAME
@@ -280,7 +280,7 @@ There are a number of numeric formats.
 .It Binary Ta Li % , 0b , 0B Ta 01
 .It Fixed-point Ta none Ta 01234.56789
 .It Precise fixed-point Ta none Ta 12.34q8
-.It Character constant Ta none Ta \(dqABYZ\(dq
+.It Character constant Ta none Ta 'ABYZ'
 .It Game Boy graphics Ta Li \` Ta 0123
 .El
 .Pp
@@ -293,11 +293,14 @@ or
 The "character constant" form yields the value the character maps to in the current charmap.
 For example, by default
 .Pq refer to Xr ascii 7
-.Sq \(dqA\(dq
+.Sq 'A'
 yields 65.
+A character constant must represent a single value, so it cannot include multiple characters, or characters which map to multiple values.
 See
 .Sx Character maps
-for information on charmaps.
+for information on charmaps, and
+.Sx String expressions
+for information on escape characters allowed in character constants.
 .Pp
 The last one, Game Boy graphics, is quite interesting and useful.
 After the backtick, 8 digits between 0 and 3 are expected, corresponding to pixel values.
@@ -538,7 +541,8 @@ There are a number of escape sequences you can use within a string:
 .Bl -column -offset indent "Sequence"
 .It Sy Sequence Ta Sy Meaning
 .It Ql \e\e Ta Backslash Pq escapes the escape character itself
-.It Ql \e" Ta Double quote Pq does not terminate the string
+.It Ql \e" Ta Double quote Pq does not terminate a string
+.It Ql \e' Ta Single quote Pq does not terminate a character literal
 .It Ql \e{ Ta Open curly brace Pq does not start interpolation
 .It Ql \e} Ta Close curly brace Pq does not end interpolation
 .It Ql \en Ta Newline Pq ASCII $0A
@@ -595,6 +599,7 @@ with its corresponding argument in
 .Pq So %% Sc is replaced by the So % Sc character .
 .It Fn STRCHAR str idx Ta Returns the substring of Ar str No for the charmap entry at Ar idx No with the current charmap . Pq Ar idx No counts charmap entries, not characters.
 .It Fn REVCHAR vals... Ta Returns the string that is mapped to Ar vals No with the current charmap. If there is no unique charmap entry for Ar vals Ns , an error occurs.
+.It Fn READFILE name max Ta Returns the contents of the file Ar name No as a string. Reads up to Ar max No bytes, or the entire contents if Ar max No is not specified. If the file isn't found in the current directory, the include-path list passed to Xr rgbasm 1 Ap s Fl I No option on the command line will be searched.
 .El
 .Pp
 The following functions operate on string expressions, but return integers.
@@ -604,6 +609,8 @@ The following functions operate on string expressions, but return integers.
 .It Fn STRCMP str1 str2 Ta Compares Ar str1 No and Ar str2 No according to ASCII ordering of their characters. Returns -1 if Ar str1 No is lower than Ar str2 Ns , 1 if Ar str1 No is greater than Ar str2 Ns , or 0 if they match.
 .It Fn STRFIND str sub Ta Returns the first index of Ar sub No in Ar str Ns , or -1 if it's not present.
 .It Fn STRRFIND str sub Ta Returns the last index of Ar sub No in Ar str Ns , or -1 if it's not present.
+.It Fn BYTELEN str Ta Returns the number of bytes in Ar str . Pq Non-ASCII characters can be multiple bytes.
+.It Fn STRBYTE str idx Ta Returns the byte value at Ar idx No in Ar str .
 .It Fn INCHARMAP str Ta Returns 1 if Ar str No has an entry in the current charmap, or 0 otherwise.
 .It Fn CHARLEN str Ta Returns the number of charmap entries in Ar str No with the current charmap.
 .It Fn CHARCMP str1 str2 Ta Compares Ar str1 No and Ar str2 No according to their charmap entry values with the current charmap. Returns -1 if Ar str1 No is lower than Ar str2 Ns , 1 if Ar str1 No is greater than Ar str2 Ns , or 0 if they match.
@@ -710,6 +717,9 @@ If
 .Ar arg
 is a section type keyword, it returns the size of that section type.
 The result is not constant, since only RGBLINK can compute its value.
+If
+.Ar arg
+is an 8-bit or 16-bit register, it returns the size of that register.
 .It Fn STARTOF arg Ta If
 .Ar arg
 is a string, this function returns the starting address of the section named
@@ -1112,6 +1122,120 @@ first, followed by the one from
 and the one from
 .Ql bar.o
 last.
+.Ss Fragment literals
+Fragment literals are useful for short blocks of code or data that are only referenced once.
+They are section fragments created by surrounding instructions or directives with
+.Ql [[
+double brackets
+.Ql ]] ,
+without a separate
+.Ic SECTION FRAGMENT
+declaration.
+.Pp
+The content of a fragment literal becomes a
+.Ic SECTION FRAGMENT ,
+sharing the same name and bank as its parent ROM section, but without any other constraints.
+The parent section also becomes a
+.Ic FRAGMENT
+if it was not one already, so that it can be merged with its fragment literals.
+RGBLINK merges the fragments in no particular order.
+.Pp
+A fragment literal can take the place of any 16-bit integer constant
+.Ql n16
+from the
+.Xr gbz80 7
+documentation, as well as a
+.Ic DW
+item.
+The fragment literal then evaluates to its starting address.
+For example, you can
+.Ic CALL
+or
+.Ic JP
+to a fragment literal.
+.Pp
+This code using named labels:
+.Bd -literal -offset indent
+DataTable:
+    dw First
+    dw Second
+    dw Third
+First:  db 1
+Second: db 4
+Third:  db 9
+Routine:
+    push hl
+    ld hl, Left
+    jr z, .got_it
+    ld hl, Right
+\&.got_it
+    call .print
+    pop hl
+    ret
+\&.print:
+    ld de, $1003
+    ld bc, STARTOF(VRAM)
+    jp Print
+Left:  db "left\e0"
+Right: db "right\e0"
+.Ed
+.Pp
+is equivalent to this code using fragment literals:
+.Bd -literal -offset indent
+DataTable:
+    dw [[ db 1 ]]
+    dw [[ db 4 ]]
+    dw [[ db 9 ]]
+Routine:
+    push hl
+    ld hl, [[ db "left\e0" ]]
+    jr z, .got_it
+    ld hl, [[ db "right\e0" ]]
+\&.got_it
+    call [[
+        ld de, $1003
+        ld bc, STARTOF(VRAM)
+        jp Print
+    ]]
+    pop hl
+    ret
+.Ed
+.Pp
+The difference is that the example using fragment literals does not declare a particular order for its pieces.
+.Pp
+Fragment literals can be arbitrarily nested, so extreme use cases are
+.Em technically
+possible.
+This code using named labels:
+.Bd -literal -offset indent
+dw FortyTwo
+FortyTwo:
+    call Sub1
+    jr Sub2
+Sub1:
+    ld a, [Twenty]
+    ret
+Twenty: db 20
+Sub2:
+    jp Sub3
+Sub3:
+    call Sub1
+    inc a
+    add a
+    ret
+.Ed
+.Pp
+is equivalent to this code using fragment literals:
+.Bd -literal -offset indent
+dw [[
+    call [[
+        Sub1: ld a, [ [[db 20]] ] :: ret
+    ]]
+    jr [[
+        jp [[ call Sub1 :: inc a :: add a :: ret ]]
+    ]]
+]]
+.Ed
 .Sh SYMBOLS
 RGBDS supports several types of symbols:
 .Bl -hang
@@ -1691,10 +1815,9 @@ Use
 .Ic INCBIN
 to include a raw binary file as it is.
 If the file isn't found in the current directory, the include-path list passed to
-.Xr rgbasm 1
-(see the
+.Xr rgbasm 1 Ap s
 .Fl I
-option) on the command line will be searched.
+option on the command line will be searched.
 .Bd -literal -offset indent
 INCBIN "titlepic.bin"
 INCBIN "sprites/hero.bin"
@@ -2247,11 +2370,10 @@ block, all of them but the first one are ignored.
 Use
 .Ic INCLUDE
 to process another assembler file and then return to the current file when done.
-If the file isn't found in the current directory, the include path list (see the
+If the file isn't found in the current directory, the include-path list passed to
+.Xr rgbasm 1 Ap s
 .Fl I
-option in
-.Xr rgbasm 1 )
-will be searched.
+option on the command line will be searched.
 You may nest
 .Ic INCLUDE
 calls infinitely (or until you run out of memory, whichever comes first).

man/rgbds.5

@@ -1,6 +1,6 @@
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBDS 5
 .Os
 .Sh NAME

man/rgbds.7

@@ -1,6 +1,6 @@
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBDS 7
 .Os
 .Sh NAME
@@ -57,7 +57,7 @@ to this day.
 .It
 2015-01-18:
 .An stag019
-begins implementing RGBGFX, a PNG‐to‐Game Boy graphics converter, for eventual integration into RGBDS.
+begins implementing RGBGFX, a PNG-to-Game Boy graphics converter, for eventual integration into RGBDS.
 .It
 2016-09-05: RGBGFX is integrated into Bentley's repository.
 .It

man/rgbfix.1

@@ -1,6 +1,6 @@
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBFIX 1
 .Os
 .Sh NAME
@@ -8,7 +8,7 @@
 .Nd Game Boy header utility and checksum fixer
 .Sh SYNOPSIS
 .Nm
-.Op Fl hjOsVv
+.Op Fl hjOsVvw
 .Op Fl C | c
 .Op Fl f Ar fix_spec
 .Op Fl i Ar game_id
@@ -21,6 +21,7 @@
 .Op Fl p Ar pad_value
 .Op Fl r Ar ram_size
 .Op Fl t Ar title_str
+.Op Fl W Ar warning
 .Op Ar
 .Sh DESCRIPTION
 The
@@ -98,7 +99,7 @@ Print help text for the program and exit.
 Set the game ID string
 .Pq Ad 0x13F Ns \(en Ns Ad 0x142
 to a given string.
-If it's longer than 4 chars, it will be truncated, and a warning emitted.
+If it's longer than 4 characters, it will be truncated.
 .It Fl j , Fl \-non-japanese
 Set the non-Japanese region flag
 .Pq Ad 0x14A
@@ -107,7 +108,7 @@ to 0x01.
 Set the new licensee string
 .Pq Ad 0x144 Ns \(en Ns Ad 0x145
 to a given string.
-If it's longer than 2 chars, it will be truncated, and a warning emitted.
+If it's longer than 2 characters, it will be truncated.
 .It Fl L Ar logo_file , Fl \-logo Ar logo_file
 Specify a logo file to use instead of the official Nintendo logo.
 The file must be 48 bytes of 1bpp tile data; the source image should be 48 pixels wide and 8 pixels tall.
@@ -124,6 +125,8 @@ to a given value from 0 to 0xFF.
 This value may also be an MBC name.
 The list of accepted names can be obtained by passing
 .Ql Cm help
+or
+.Ql Cm list
 as the argument.
 Any amount of whitespace (space and tabs) is allowed around plus signs, and the order of "components" is free, as long as the MBC name is first.
 There are special considerations to take for the TPP1 mapper; see the
@@ -134,7 +137,8 @@ Set the ROM version
 .Pq Ad 0x14C
 to a given value from 0 to 0xFF.
 .It Fl O , Fl \-overwrite
-Allow overwriting different non-zero bytes in the header without a warning being emitted.
+Alias for
+.Fl Wno-overwrite .
 .It Fl o Ar out_file , Fl \-output Ar out_file
 Write the modified ROM image to the given file, or '-' to write to standard output.
 If not specified, the input files are modified in-place, or written to standard output if read from standard input.
@@ -154,15 +158,14 @@ to a given value from 0 to 0xFF.
 Set the SGB flag
 .Pq Ad 0x146
 to 0x03.
-This flag will be ignored by the SGB unless the old licensee code is 0x33!
-If this is given as well as
-.Fl l ,
-but is not set to 0x33, a warning will be printed.
+This flag will be ignored by the SGB unless the old licensee code
+.Pq Fl -l
+is 0x33!
 .It Fl t Ar title , Fl \-title Ar title
 Set the title string
 .Pq Ad 0x134 Ns \(en Ns Ad 0x143
 to a given string.
-If the title is longer than the max length, it will be truncated, and a warning emitted.
+If the title is longer than the maximum length, it will be truncated.
 The max length is 11 characters if the game ID
 .Pq Fl i
 is specified, 15 characters if the CGB flag
@@ -175,6 +178,77 @@ Print the version of the program and exit.
 .It Fl v , Fl \-validate
 Equivalent to
 .Fl f Cm lhg .
+.It Fl W Ar warning , Fl \-warning Ar warning
+Set warning flag
+.Ar warning .
+A warning message will be printed if
+.Ar warning
+is an unknown warning flag.
+See the
+.Sx DIAGNOSTICS
+section for a list of warnings.
+.It Fl w
+Disable all warning output, even when turned into errors.
+.El
+.Sh DIAGNOSTICS
+Warnings are diagnostic messages that indicate possibly erroneous behavior that does not necessarily compromise the header-fixing process.
+The following options alter the way warnings are processed.
+.Bl -tag -width Ds
+.It Fl Werror
+Make all warnings into errors.
+This can be negated as
+.Fl Wno-error
+to prevent turning all warnings into errors.
+.It Fl Werror=
+Make the specified warning or meta warning into an error.
+A warning's name is appended
+.Pq example: Fl Werror=overwrite ,
+and this warning is implicitly enabled and turned into an error.
+This can be negated as
+.Fl Wno-error=
+to prevent turning a specified warning into an error, even if
+.Fl Werror
+is in effect.
+.El
+.Pp
+The following warnings are
+.Dq meta
+warnings, that enable a collection of other warnings.
+If a specific warning is toggled via a meta flag and a specific one, the more specific one takes priority.
+The position on the command-line acts as a tie breaker, the last one taking effect.
+.Bl -tag -width Ds
+.It Fl Wall
+This enables warnings that are likely to indicate an error or undesired behavior, and that can easily be fixed.
+.It Fl Weverything
+Enables literally every warning.
+.El
+.Pp
+The following warnings are actual warning flags; with each description, the corresponding warning flag is included.
+Note that each of these flag also has a negation (for example,
+.Fl Wtruncation
+enables the warning that
+.Fl Wno-truncation
+disables; and
+.Fl Wall
+enables every warning that
+.Fl Wno-all
+disables).
+Only the non-default flag is listed here.
+Ignoring the
+.Dq no-
+prefix, entries are listed alphabetically.
+.Bl -tag -width Ds
+.It Fl Wno-mbc
+Warn when there are inconsistencies with or caveats about the specified MBC type.
+.It Fl Wno-overwrite
+Warn when overwriting different non-zero bytes in the header.
+.It Fl Wno-sgb
+Warn when the SGB flag
+.Pq Fl s
+conflicts with the old licensee code
+.Pq Fl l .
+.It Fl Wno-truncation
+Warn when truncating values to fit the available space.
 .El
 .Sh EXAMPLES
 Most values in the ROM header do not matter to the actual console, and most are seldom useful anyway.
@@ -228,7 +302,7 @@ Therefore,
 .Nm
 will ignore the
 .Ql RAM
-feature on a TPP1 mapper with a warning.
+feature on a TPP1 mapper.
 .Ss Special considerations
 TPP1 overwrites the byte at
 .Ad 0x14A ,

man/rgbgfx.1

@@ -2,7 +2,7 @@
 .\"
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBGFX 1
 .Os
 .Sh NAME
@@ -10,7 +10,7 @@
 .Nd Game Boy graphics converter
 .Sh SYNOPSIS
 .Nm
-.Op Fl CmhOuVXYZ
+.Op Fl CmhOuVwXYZ
 .Op Fl v Op Fl v No ...
 .Op Fl a Ar attrmap | Fl A
 .Op Fl b Ar base_ids
@@ -18,6 +18,7 @@
 .Op Fl d Ar depth
 .Op Fl i Ar input_tiles
 .Op Fl L Ar slice
+.Op Fl l Ar base_pal
 .Op Fl N Ar nb_tiles
 .Op Fl n Ar nb_pals
 .Op Fl o Ar out_file
@@ -26,6 +27,7 @@
 .Op Fl r Ar width
 .Op Fl s Ar nb_colors
 .Op Fl t Ar tilemap | Fl T
+.Op Fl W Ar warning
 .Op Fl x Ar quantity
 .Ar file
 .Sh DESCRIPTION
@@ -261,6 +263,11 @@ The first number pair specifies the X and Y coordinates of the top-left pixel th
 The second number pair specifies how many tiles to process horizontally and vertically, respectively.
 .Pp
 .Fl L Sy is ignored in reverse mode , No no padding is inserted .
+.It Fl l Ar base_pal , Fl \-base-palette Ar base_pal
+Set the base ID for attribute map and palette map output.
+.Ar base_pal
+should be a number between 0 and 255.
+It defaults to 0.
 .It Fl m , Fl \-mirror-tiles
 Deduplicate tiles that are horizontally and/or vertically symmetrical mirror images of each other.
 Only one of each unique tile will be saved in the tile data file, with mirror images counting as duplicates.
@@ -387,6 +394,17 @@ Some internal debug printing is enabled.
 The verbosity level does not go past 6.
 .Pp
 Note that verbose output is only intended to be consumed by humans, and may change without notice between RGBDS releases; relying on those for scripts is not advised.
+.It Fl W Ar warning , Fl \-warning Ar warning
+Set warning flag
+.Ar warning .
+A warning message will be printed if
+.Ar warning
+is an unknown warning flag.
+See the
+.Sx DIAGNOSTICS
+section for a list of warnings.
+.It Fl w
+Disable all warning output, even when turned into errors.
 .It Fl X , Fl \-mirror-x
 Deduplicate tiles that are horizontally symmetrical mirror images of each other across the X axis.
 Implies
@@ -498,6 +516,9 @@ Useful to force several images to share the same palette.
 Plaintext lines of hexadecimal colors in
 .Ql rrggbb
 format.
+.It Cm png
+An image of square color swatches, with each row defining the colors for one palette.
+Color swatches can be any square size.
 .It Cm psp
 .Lk https://www.selapa.net/swatches/colors/fileformats.php#psp_pal Paint Shop Pro palette .
 .El
@@ -702,6 +723,68 @@ assume that tiles were not deduplicated, and should be laid out in the order the
 .Nm
 assumes that no tiles were mirrored.
 .El
+.Sh DIAGNOSTICS
+Warnings are diagnostic messages that indicate possibly erroneous behavior that does not necessarily compromise the conversion process.
+The following options alter the way warnings are processed.
+.Bl -tag -width Ds
+.It Fl Werror
+Make all warnings into errors.
+This can be negated as
+.Fl Wno-error
+to prevent turning all warnings into errors.
+.It Fl Werror=
+Make the specified warning or meta warning into an error.
+A warning's name is appended
+.Pq example: Fl Werror=embedded ,
+and this warning is implicitly enabled and turned into an error.
+This can be negated as
+.Fl Wno-error=
+to prevent turning a specified warning into an error, even if
+.Fl Werror
+is in effect.
+.El
+.Pp
+The following warnings are
+.Dq meta
+warnings, that enable a collection of other warnings.
+If a specific warning is toggled via a meta flag and a specific one, the more specific one takes priority.
+The position on the command-line acts as a tie breaker, the last one taking effect.
+.Bl -tag -width Ds
+.It Fl Wall
+This enables warnings that are likely to indicate an error or undesired behavior, and that can easily be fixed.
+.It Fl Weverything
+Enables literally every warning.
+.El
+.Pp
+The following warnings are actual warning flags; with each description, the corresponding warning flag is included.
+Note that each of these flag also has a negation (for example,
+.Fl Wtrim-nonempty
+enables the warning that
+.Fl Wno-trim-nonempty
+disables; and
+.Fl Wall
+enables every warning that
+.Fl Wno-all
+disables).
+Only the non-default flag is listed here.
+Ignoring the
+.Dq no-
+prefix, entries are listed alphabetically.
+.Bl -tag -width Ds
+.It Fl Wembedded
+Warn when a generated palette is sorted according to the input PNG's embedded palette but
+.Fl c Cm embedded
+was not provided.
+This warning is enabled by
+.Fl Weverything .
+.It Fl Wtrim-nonempty
+Warn when
+.Fl x
+trims a nonempty tile.
+An "empty" tile uses entirely color 0 of its palette.
+This warning is enabled by
+.Fl Wall .
+.El
 .Sh EXAMPLES
 The following will only validate the
 .Ql tileset.png

man/rgblink.1

@@ -1,6 +1,6 @@
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBLINK 1
 .Os
 .Sh NAME
@@ -16,6 +16,7 @@
 .Op Fl o Ar out_file
 .Op Fl p Ar pad_value
 .Op Fl S Ar spec
+.Op Fl W Ar warning
 .Ar
 .Sh DESCRIPTION
 The
@@ -114,6 +115,15 @@ Useful for ROMs that fit in 32 KiB.
 Print the version of the program and exit.
 .It Fl v , Fl \-verbose
 Verbose: enable printing more information to standard error.
+.It Fl W Ar warning , Fl \-warning Ar warning
+Set warning flag
+.Ar warning .
+A warning message will be printed if
+.Ar warning
+is an unknown warning flag.
+See the
+.Sx DIAGNOSTICS
+section for a list of warnings.
 .It Fl w , Fl \-wramx
 Expand the WRAM0 section size from 4 KiB to the full 8 KiB assigned to WRAM.
 WRAMX sections that are fixed to a bank other than 1 become errors, other WRAMX sections are treated as WRAM0.
@@ -176,6 +186,84 @@ as
 .Ic WRAMX
 sections will be treated as
 .Ic WRAM0 .
+.Sh DIAGNOSTICS
+Warnings are diagnostic messages that indicate possibly erroneous behavior that does not necessarily compromise the linking process.
+The following options alter the way warnings are processed.
+.Bl -tag -width Ds
+.It Fl Werror
+Make all warnings into errors.
+This can be negated as
+.Fl Wno-error
+to prevent turning all warnings into errors.
+.It Fl Werror=
+Make the specified warning or meta warning into an error.
+A warning's name is appended
+.Pq example: Fl Werror=assert ,
+and this warning is implicitly enabled and turned into an error.
+This can be negated as
+.Fl Wno-error=
+to prevent turning a specified warning into an error, even if
+.Fl Werror
+is in effect.
+.El
+.Pp
+The following warnings are
+.Dq meta
+warnings, that enable a collection of other warnings.
+If a specific warning is toggled via a meta flag and a specific one, the more specific one takes priority.
+The position on the command-line acts as a tie breaker, the last one taking effect.
+.Bl -tag -width Ds
+.It Fl Wall
+This enables warnings that are likely to indicate an error or undesired behavior, and that can easily be fixed.
+.It Fl Weverything
+Enables literally every warning.
+.El
+.Pp
+The following warnings are actual warning flags; with each description, the corresponding warning flag is included.
+Note that each of these flag also has a negation (for example,
+.Fl Wobsolete
+enables the warning that
+.Fl Wno-obsolete
+disables; and
+.Fl Wall
+enables every warning that
+.Fl Wno-all
+disables).
+Only the non-default flag is listed here.
+Ignoring the
+.Dq no-
+prefix, entries are listed alphabetically.
+.Bl -tag -width Ds
+.It Fl Wno-assert
+Warn when
+.Ic WARN Ns No -type
+assertions fail. (See
+.Dq Aborting the assembly process
+in
+.Xr rgbasm 5
+for
+.Ic ASSERT ) .
+.It Fl Wdiv
+Warn when dividing the smallest negative integer (-2**31) by -1, which yields itself due to integer overflow.
+This warning is enabled by
+.Fl Wall .
+.It Fl Wno-obsolete
+Warn when obsolete features are encountered, which have been deprecated and may later be removed.
+.It Fl Wshift
+Warn when shifting right a negative value.
+Use a division by 2**N instead.
+This warning is enabled by
+.Fl Wall .
+.It Fl Wshift-amount
+Warn when a shift's operand is negative or greater than 32.
+This warning is enabled by
+.Fl Wall .
+.It Fl Wno-truncation
+Warn when an implicit truncation (for example,
+.Ic db
+to an 8-bit value) loses some bits.
+This occurs when an N-bit value is 2**N or greater, or less than -2**N.
+.El
 .Sh EXAMPLES
 All you need for a basic ROM is an object file, which can be made into a ROM image like so:
 .Pp

man/rgblink.5

@@ -1,6 +1,6 @@
 .\" SPDX-License-Identifier: MIT
 .\"
-.Dd June 30, 2025
+.Dd July 31, 2025
 .Dt RGBLINK 5
 .Os
 .Sh NAME
@@ -24,18 +24,20 @@ They are simply ignored.
 .Pp
 Keywords are composed of letters and digits (but they can't start with a digit); they are all case-insensitive.
 .Pp
-Numbers can be written in decimal format, or in binary using the
-.Ql %
-prefix, or in hexadecimal using the
-.Ql $
-prefix (hexadecimal digits are case-insensitive).
-Note that unlike
-.Xr rgbasm 5 ,
-an octal
-.Ql &
-prefix is not supported, nor are
-.Ql _
-digit separators.
+Numbers can be written in a number of formats.
+.Bl -column -offset indent "Hexadecimal" "Possible prefixes"
+.It Sy Format type Ta Sy Possible prefixes Ta Sy Accepted characters
+.It Decimal Ta none Ta 0123456789
+.It Hexadecimal Ta Li $ , 0x , 0X Ta 0123456789ABCDEF
+.It Octal Ta Li & , 0o , 0O Ta 01234567
+.It Binary Ta Li % , 0b , 0B Ta 01
+.El
+.Pp
+Underscores are also accepted in numbers, except at the beginning of one.
+This can be useful for grouping digits, like
+.Ql 1_234
+or
+.Ql $ff_80 .
 .Pp
 Strings begin with a double quote, and end at the next (non-escaped) double quote.
 Strings must not contain literal newline characters.
@@ -46,8 +48,9 @@ are supported, specifically
 .Ql \e" ,
 .Ql \en ,
 .Ql \er ,
+.Ql \et ,
 and
-.Ql \et .
+.Ql \e0 .
 Other backslash escape sequences in
 .Xr rgbasm 5
 are only relevant to assembly code and do not apply in linker scripts.

src/CMakeLists.txt

@@ -3,8 +3,9 @@
 configure_file(version.cpp _version.cpp ESCAPE_QUOTES)
 
 set(common_src
-    "error.cpp"
     "extern/getopt.cpp"
+    "diagnostics.cpp"
+    "usage.cpp"
     "_version.cpp"
     )
 
@@ -34,6 +35,7 @@ BISON_TARGET(LINKER_SCRIPT_PARSER "link/script.y"
 
 set(rgbasm_src
     "${BISON_ASM_PARSER_OUTPUT_SOURCE}"
+    "asm/actions.cpp"
     "asm/charmap.cpp"
     "asm/fixpoint.cpp"
     "asm/format.cpp"
@@ -53,26 +55,11 @@ set(rgbasm_src
     "util.cpp"
     )
 
-set(rgbfix_src
-    "fix/main.cpp"
-    )
-
-set(rgbgfx_src
-    "gfx/main.cpp"
-    "gfx/pal_packing.cpp"
-    "gfx/pal_sorting.cpp"
-    "gfx/pal_spec.cpp"
-    "gfx/process.cpp"
-    "gfx/proto_palette.cpp"
-    "gfx/reverse.cpp"
-    "gfx/rgba.cpp"
-    "extern/getopt.cpp"
-    "error.cpp"
-    )
-
 set(rgblink_src
     "${BISON_LINKER_SCRIPT_PARSER_OUTPUT_SOURCE}"
     "link/assign.cpp"
+    "link/lexer.cpp"
+    "link/layout.cpp"
     "link/main.cpp"
     "link/object.cpp"
     "link/output.cpp"
@@ -80,12 +67,33 @@ set(rgblink_src
     "link/sdas_obj.cpp"
     "link/section.cpp"
     "link/symbol.cpp"
+    "link/warning.cpp"
     "extern/utf8decoder.cpp"
     "linkdefs.cpp"
     "opmath.cpp"
     "util.cpp"
     )
 
+set(rgbfix_src
+    "fix/main.cpp"
+    "fix/mbc.cpp"
+    "fix/warning.cpp"
+    )
+
+set(rgbgfx_src
+    "gfx/color_set.cpp"
+    "gfx/main.cpp"
+    "gfx/pal_packing.cpp"
+    "gfx/pal_sorting.cpp"
+    "gfx/pal_spec.cpp"
+    "gfx/png.cpp"
+    "gfx/process.cpp"
+    "gfx/reverse.cpp"
+    "gfx/rgba.cpp"
+    "gfx/warning.cpp"
+    "util.cpp"
+    )
+
 foreach(PROG "asm" "fix" "gfx" "link")
   add_executable(rgb${PROG}
                  ${rgb${PROG}_src}

src/asm/actions.cpp

@@ -0,0 +1,469 @@
+// SPDX-License-Identifier: MIT
+
+#include "asm/actions.hpp"
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "extern/utf8decoder.hpp"
+#include "helpers.hpp"
+
+#include "asm/charmap.hpp"
+#include "asm/format.hpp"
+#include "asm/fstack.hpp"
+#include "asm/symbol.hpp"
+#include "asm/warning.hpp"
+
+std::optional<std::string> act_ReadFile(std::string const &name, uint32_t maxLen) {
+	FILE *file = nullptr;
+	if (std::optional<std::string> fullPath = fstk_FindFile(name); fullPath) {
+		file = fopen(fullPath->c_str(), "rb");
+	}
+	if (!file) {
+		if (fstk_FileError(name, "READFILE")) {
+			// If `fstk_FileError` returned true due to `-MG`, we should abort due to a
+			// missing file, so return `std::nullopt`, which tells the caller to `YYACCEPT`
+			return std::nullopt;
+		}
+		return "";
+	}
+	Defer closeFile{[&] { fclose(file); }};
+
+	size_t readSize = maxLen;
+	if (fseek(file, 0, SEEK_END) == 0) {
+		// If the file is seekable and shorter than the max length,
+		// just read as many bytes as there are
+		if (long fileSize = ftell(file); static_cast<size_t>(fileSize) < readSize) {
+			readSize = fileSize;
+		}
+		fseek(file, 0, SEEK_SET);
+	} else if (errno != ESPIPE) {
+		error("Error determining size of READFILE file '%s': %s", name.c_str(), strerror(errno));
+	}
+
+	std::string contents;
+	contents.resize(readSize);
+
+	if (fread(&contents[0], 1, readSize, file) < readSize || ferror(file)) {
+		error("Error reading READFILE file '%s': %s", name.c_str(), strerror(errno));
+		return "";
+	}
+
+	return contents;
+}
+
+uint32_t act_StringToNum(std::vector<int32_t> const &str) {
+	uint32_t length = str.size();
+
+	if (length == 1) {
+		// The string is a single character with a single value,
+		// which can be used directly as a number.
+		return static_cast<uint32_t>(str[0]);
+	}
+
+	warning(WARNING_OBSOLETE, "Treating multi-unit strings as numbers is deprecated");
+
+	for (int32_t v : str) {
+		if (!checkNBit(v, 8, "All character units")) {
+			break;
+		}
+	}
+
+	uint32_t r = 0;
+
+	for (uint32_t i = length < 4 ? 0 : length - 4; i < length; ++i) {
+		r <<= 8;
+		r |= static_cast<uint8_t>(str[i]);
+	}
+
+	return r;
+}
+
+static void errorInvalidUTF8Byte(uint8_t byte, char const *functionName) {
+	error("%s: Invalid UTF-8 byte 0x%02hhX", functionName, byte);
+}
+
+size_t act_StringLen(std::string const &str, bool printErrors) {
+	size_t len = 0;
+	uint32_t state = UTF8_ACCEPT;
+	uint32_t codepoint = 0;
+
+	for (char c : str) {
+		uint8_t byte = static_cast<uint8_t>(c);
+
+		switch (decode(&state, &codepoint, byte)) {
+		case UTF8_REJECT:
+			if (printErrors) {
+				errorInvalidUTF8Byte(byte, "STRLEN");
+			}
+			state = UTF8_ACCEPT;
+			// fallthrough
+		case UTF8_ACCEPT:
+			++len;
+			break;
+		}
+	}
+
+	// Check for partial code point.
+	if (state != UTF8_ACCEPT) {
+		if (printErrors) {
+			error("STRLEN: Incomplete UTF-8 character");
+		}
+		++len;
+	}
+
+	return len;
+}
+
+std::string act_StringSlice(std::string const &str, uint32_t start, uint32_t stop) {
+	size_t strLen = str.length();
+	size_t index = 0;
+	uint32_t state = UTF8_ACCEPT;
+	uint32_t codepoint = 0;
+	uint32_t curIdx = 0;
+
+	// Advance to starting index in source string.
+	while (index < strLen && curIdx < start) {
+		switch (decode(&state, &codepoint, str[index])) {
+		case UTF8_REJECT:
+			errorInvalidUTF8Byte(str[index], "STRSLICE");
+			state = UTF8_ACCEPT;
+			// fallthrough
+		case UTF8_ACCEPT:
+			++curIdx;
+			break;
+		}
+		++index;
+	}
+
+	// An index 1 past the end of the string is allowed, but will trigger the
+	// "Length too big" warning below if the length is nonzero.
+	if (index >= strLen && start > curIdx) {
+		warning(
+		    WARNING_BUILTIN_ARG,
+		    "STRSLICE: Start index %" PRIu32 " is past the end of the string",
+		    start
+		);
+	}
+
+	size_t startIndex = index;
+
+	// Advance to ending index in source string.
+	while (index < strLen && curIdx < stop) {
+		switch (decode(&state, &codepoint, str[index])) {
+		case UTF8_REJECT:
+			errorInvalidUTF8Byte(str[index], "STRSLICE");
+			state = UTF8_ACCEPT;
+			// fallthrough
+		case UTF8_ACCEPT:
+			++curIdx;
+			break;
+		}
+		++index;
+	}
+
+	// Check for partial code point.
+	if (state != UTF8_ACCEPT) {
+		error("STRSLICE: Incomplete UTF-8 character");
+		++curIdx;
+	}
+
+	if (curIdx < stop) {
+		warning(
+		    WARNING_BUILTIN_ARG,
+		    "STRSLICE: Stop index %" PRIu32 " is past the end of the string",
+		    stop
+		);
+	}
+
+	return str.substr(startIndex, index - startIndex);
+}
+
+std::string act_StringSub(std::string const &str, uint32_t pos, uint32_t len) {
+	size_t strLen = str.length();
+	size_t index = 0;
+	uint32_t state = UTF8_ACCEPT;
+	uint32_t codepoint = 0;
+	uint32_t curPos = 1;
+
+	// Advance to starting position in source string.
+	while (index < strLen && curPos < pos) {
+		switch (decode(&state, &codepoint, str[index])) {
+		case UTF8_REJECT:
+			errorInvalidUTF8Byte(str[index], "STRSUB");
+			state = UTF8_ACCEPT;
+			// fallthrough
+		case UTF8_ACCEPT:
+			++curPos;
+			break;
+		}
+		++index;
+	}
+
+	// A position 1 past the end of the string is allowed, but will trigger the
+	// "Length too big" warning below if the length is nonzero.
+	if (index >= strLen && pos > curPos) {
+		warning(
+		    WARNING_BUILTIN_ARG, "STRSUB: Position %" PRIu32 " is past the end of the string", pos
+		);
+	}
+
+	size_t startIndex = index;
+	uint32_t curLen = 0;
+
+	// Compute the result length in bytes.
+	while (index < strLen && curLen < len) {
+		switch (decode(&state, &codepoint, str[index])) {
+		case UTF8_REJECT:
+			errorInvalidUTF8Byte(str[index], "STRSUB");
+			state = UTF8_ACCEPT;
+			// fallthrough
+		case UTF8_ACCEPT:
+			++curLen;
+			break;
+		}
+		++index;
+	}
+
+	// Check for partial code point.
+	if (state != UTF8_ACCEPT) {
+		error("STRSUB: Incomplete UTF-8 character");
+		++curLen;
+	}
+
+	if (curLen < len) {
+		warning(WARNING_BUILTIN_ARG, "STRSUB: Length too big: %" PRIu32, len);
+	}
+
+	return str.substr(startIndex, index - startIndex);
+}
+
+size_t act_CharLen(std::string const &str) {
+	std::string_view view = str;
+	size_t len;
+
+	for (len = 0; charmap_ConvertNext(view, nullptr); ++len) {}
+
+	return len;
+}
+
+std::string act_StringChar(std::string const &str, uint32_t idx) {
+	std::string_view view = str;
+	size_t charLen = 1;
+
+	// Advance to starting index in source string.
+	for (uint32_t curIdx = 0; charLen && curIdx < idx; ++curIdx) {
+		charLen = charmap_ConvertNext(view, nullptr);
+	}
+
+	std::string_view start = view;
+
+	if (!charmap_ConvertNext(view, nullptr)) {
+		warning(
+		    WARNING_BUILTIN_ARG, "STRCHAR: Index %" PRIu32 " is past the end of the string", idx
+		);
+	}
+
+	start = start.substr(0, start.length() - view.length());
+	return std::string(start);
+}
+
+std::string act_CharSub(std::string const &str, uint32_t pos) {
+	std::string_view view = str;
+	size_t charLen = 1;
+
+	// Advance to starting position in source string.
+	for (uint32_t curPos = 1; charLen && curPos < pos; ++curPos) {
+		charLen = charmap_ConvertNext(view, nullptr);
+	}
+
+	std::string_view start = view;
+
+	if (!charmap_ConvertNext(view, nullptr)) {
+		warning(
+		    WARNING_BUILTIN_ARG, "CHARSUB: Position %" PRIu32 " is past the end of the string", pos
+		);
+	}
+
+	start = start.substr(0, start.length() - view.length());
+	return std::string(start);
+}
+
+int32_t act_CharCmp(std::string_view str1, std::string_view str2) {
+	std::vector<int32_t> seq1, seq2;
+	size_t idx1 = 0, idx2 = 0;
+	for (;;) {
+		if (idx1 >= seq1.size()) {
+			idx1 = 0;
+			seq1.clear();
+			charmap_ConvertNext(str1, &seq1);
+		}
+		if (idx2 >= seq2.size()) {
+			idx2 = 0;
+			seq2.clear();
+			charmap_ConvertNext(str2, &seq2);
+		}
+		if (seq1.empty() != seq2.empty()) {
+			return seq1.empty() ? -1 : 1;
+		} else if (seq1.empty()) {
+			return 0;
+		} else {
+			int32_t value1 = seq1[idx1++], value2 = seq2[idx2++];
+			if (value1 != value2) {
+				return (value1 > value2) - (value1 < value2);
+			}
+		}
+	}
+}
+
+uint32_t act_AdjustNegativeIndex(int32_t idx, size_t len, char const *functionName) {
+	// String functions adjust negative index arguments the same way,
+	// such that position -1 is the last character of a string.
+	if (idx < 0) {
+		idx += len;
+	}
+	if (idx < 0) {
+		warning(WARNING_BUILTIN_ARG, "%s: Index starts at 0", functionName);
+		idx = 0;
+	}
+	return static_cast<uint32_t>(idx);
+}
+
+uint32_t act_AdjustNegativePos(int32_t pos, size_t len, char const *functionName) {
+	// STRSUB and CHARSUB adjust negative position arguments the same way,
+	// such that position -1 is the last character of a string.
+	if (pos < 0) {
+		pos += len + 1;
+	}
+	if (pos < 1) {
+		warning(WARNING_BUILTIN_ARG, "%s: Position starts at 1", functionName);
+		pos = 1;
+	}
+	return static_cast<uint32_t>(pos);
+}
+
+std::string
+    act_StringReplace(std::string_view str, std::string const &old, std::string const &rep) {
+	if (old.empty()) {
+		warning(WARNING_EMPTY_STRRPL, "STRRPL: Cannot replace an empty string");
+		return std::string(str);
+	}
+
+	std::string rpl;
+
+	while (!str.empty()) {
+		auto pos = str.find(old);
+		if (pos == str.npos) {
+			rpl.append(str);
+			break;
+		}
+		rpl.append(str, 0, pos);
+		rpl.append(rep);
+		str.remove_prefix(pos + old.size());
+	}
+
+	return rpl;
+}
+
+std::string act_StringFormat(
+    std::string const &spec, std::vector<std::variant<uint32_t, std::string>> const &args
+) {
+	std::string str;
+	size_t argIndex = 0;
+
+	for (size_t i = 0; spec[i] != '\0'; ++i) {
+		int c = spec[i];
+
+		if (c != '%') {
+			str += c;
+			continue;
+		}
+
+		c = spec[++i];
+
+		if (c == '%') {
+			str += c;
+			continue;
+		}
+
+		FormatSpec fmt{};
+
+		while (c != '\0') {
+			fmt.useCharacter(c);
+			if (fmt.isFinished()) {
+				break;
+			}
+			c = spec[++i];
+		}
+
+		if (fmt.isEmpty()) {
+			error("STRFMT: Illegal '%%' at end of format string");
+			str += '%';
+			break;
+		}
+
+		if (!fmt.isValid()) {
+			error("STRFMT: Invalid format spec for argument %zu", argIndex + 1);
+			str += '%';
+		} else if (argIndex >= args.size()) {
+			// Will warn after formatting is done.
+			str += '%';
+		} else if (std::holds_alternative<uint32_t>(args[argIndex])) {
+			fmt.appendNumber(str, std::get<uint32_t>(args[argIndex]));
+		} else {
+			fmt.appendString(str, std::get<std::string>(args[argIndex]));
+		}
+
+		++argIndex;
+	}
+
+	if (argIndex < args.size()) {
+		error("STRFMT: %zu unformatted argument(s)", args.size() - argIndex);
+	} else if (argIndex > args.size()) {
+		error(
+		    "STRFMT: Not enough arguments for format spec, got: %zu, need: %zu",
+		    args.size(),
+		    argIndex
+		);
+	}
+
+	return str;
+}
+
+void act_CompoundAssignment(std::string const &symName, RPNCommand op, int32_t constValue) {
+	Expression oldExpr, constExpr, newExpr;
+	int32_t newValue;
+
+	oldExpr.makeSymbol(symName);
+	constExpr.makeNumber(constValue);
+	newExpr.makeBinaryOp(op, std::move(oldExpr), constExpr);
+	newValue = newExpr.getConstVal();
+	sym_AddVar(symName, newValue);
+}
+
+void act_FailAssert(AssertionType type) {
+	switch (type) {
+	case ASSERT_FATAL:
+		fatal("Assertion failed");
+	case ASSERT_ERROR:
+		error("Assertion failed");
+		break;
+	case ASSERT_WARN:
+		warning(WARNING_ASSERT, "Assertion failed");
+		break;
+	}
+}
+
+void act_FailAssertMsg(AssertionType type, std::string const &message) {
+	switch (type) {
+	case ASSERT_FATAL:
+		fatal("Assertion failed: %s", message.c_str());
+	case ASSERT_ERROR:
+		error("Assertion failed: %s", message.c_str());
+		break;
+	case ASSERT_WARN:
+		warning(WARNING_ASSERT, "Assertion failed: %s", message.c_str());
+		break;
+	}
+}

src/asm/charmap.cpp

@@ -31,36 +31,34 @@ struct CharmapNode {
 struct Charmap {
 	std::string name;
 	std::vector<CharmapNode> nodes; // first node is reserved for the root node
+};
 
-	// Traverse the trie depth-first to derive the character mappings in definition order
-	template<typename F>
-	bool forEachChar(F callback) const {
+// Traverse the trie depth-first to derive the character mappings in definition order
+template<typename F>
+bool forEachChar(Charmap const &charmap, F callback) {
 	// clang-format off: nested initializers
 	for (std::stack<std::pair<size_t, std::string>> prefixes({{0, ""}}); !prefixes.empty();) {
 		// clang-format on
 		auto [nodeIdx, mapping] = std::move(prefixes.top());
 		prefixes.pop();
-			CharmapNode const &node = nodes[nodeIdx];
-			if (node.isTerminal()) {
-				if (!callback(nodeIdx, mapping)) {
+		CharmapNode const &node = charmap.nodes[nodeIdx];
+		if (node.isTerminal() && !callback(nodeIdx, mapping)) {
 			return false;
 		}
-			}
-			for (unsigned c = 0; c < std::size(node.next); c++) {
+		for (unsigned c = 0; c < std::size(node.next); ++c) {
 			if (size_t nextIdx = node.next[c]; nextIdx) {
 				prefixes.push({nextIdx, mapping + static_cast<char>(c)});
 			}
 		}
 	}
 	return true;
-	}
-};
+}
 
 static std::deque<Charmap> charmapList;
 static std::unordered_map<std::string, size_t> charmapMap; // Indexes into `charmapList`
 
 static Charmap *currentCharmap;
-std::stack<Charmap *> charmapStack;
+static std::stack<Charmap *> charmapStack;
 
 bool charmap_ForEach(
     void (*mapFunc)(std::string const &),
@@ -68,7 +66,7 @@ bool charmap_ForEach(
 ) {
 	for (Charmap const &charmap : charmapList) {
 		std::map<size_t, std::string> mappings;
-		charmap.forEachChar([&mappings](size_t nodeIdx, std::string const &mapping) {
+		forEachChar(charmap, [&mappings](size_t nodeIdx, std::string const &mapping) {
 			mappings[nodeIdx] = mapping;
 			return true;
 		});
@@ -86,14 +84,14 @@ void charmap_New(std::string const &name, std::string const *baseName) {
 
 	if (baseName != nullptr) {
 		if (auto search = charmapMap.find(*baseName); search == charmapMap.end()) {
-			error("Base charmap '%s' doesn't exist\n", baseName->c_str());
+			error("Base charmap '%s' doesn't exist", baseName->c_str());
 		} else {
 			baseIdx = search->second;
 		}
 	}
 
 	if (charmapMap.find(name) != charmapMap.end()) {
-		error("Charmap '%s' already exists\n", name.c_str());
+		error("Charmap '%s' already exists", name.c_str());
 		return;
 	}
 
@@ -114,7 +112,7 @@ void charmap_New(std::string const &name, std::string const *baseName) {
 
 void charmap_Set(std::string const &name) {
 	if (auto search = charmapMap.find(name); search == charmapMap.end()) {
-		error("Charmap '%s' doesn't exist\n", name.c_str());
+		error("Charmap '%s' doesn't exist", name.c_str());
 	} else {
 		currentCharmap = &charmapList[search->second];
 	}
@@ -126,7 +124,7 @@ void charmap_Push() {
 
 void charmap_Pop() {
 	if (charmapStack.empty()) {
-		error("No entries in the charmap stack\n");
+		error("No entries in the charmap stack");
 		return;
 	}
 
@@ -136,13 +134,13 @@ void charmap_Pop() {
 
 void charmap_CheckStack() {
 	if (!charmapStack.empty()) {
-		warning(WARNING_UNMATCHED_DIRECTIVE, "`PUSHC` without corresponding `POPC`\n");
+		warning(WARNING_UNMATCHED_DIRECTIVE, "`PUSHC` without corresponding `POPC`");
 	}
 }
 
 void charmap_Add(std::string const &mapping, std::vector<int32_t> &&value) {
 	if (mapping.empty()) {
-		error("Cannot map an empty string\n");
+		error("Cannot map an empty string");
 		return;
 	}
 
@@ -168,7 +166,7 @@ void charmap_Add(std::string const &mapping, std::vector<int32_t> &&value) {
 	CharmapNode &node = charmap.nodes[nodeIdx];
 
 	if (node.isTerminal()) {
-		warning(WARNING_CHARMAP_REDEF, "Overriding charmap mapping\n");
+		warning(WARNING_CHARMAP_REDEF, "Overriding charmap mapping");
 	}
 
 	std::swap(node.value, value);
@@ -240,13 +238,13 @@ size_t charmap_ConvertNext(std::string_view &input, std::vector<int32_t> *output
 			break;
 		}
 
-		inputIdx++; // Consume that char
+		++inputIdx; // Consume that char
 
 		if (charmap.nodes[nodeIdx].isTerminal()) {
 			matchIdx = nodeIdx; // This node matches, register it
 			rewindDistance = 0; // If no longer match is found, rewind here
 		} else {
-			rewindDistance++;
+			++rewindDistance;
 		}
 	}
 
@@ -266,14 +264,15 @@ size_t charmap_ConvertNext(std::string_view &input, std::vector<int32_t> *output
 	} else if (inputIdx < input.length()) { // No match found, but there is some input left
 		size_t codepointLen = 0;
 		// This will write the codepoint's value to `output`, little-endian
-		for (uint32_t state = 0, codepoint = 0; inputIdx + codepointLen < input.length();) {
-			if (decode(&state, &codepoint, input[inputIdx + codepointLen]) == 1) {
-				error("Input string is not valid UTF-8\n");
+		for (uint32_t state = UTF8_ACCEPT, codepoint = 0;
+		     inputIdx + codepointLen < input.length();) {
+			if (decode(&state, &codepoint, input[inputIdx + codepointLen]) == UTF8_REJECT) {
+				error("Input string is not valid UTF-8");
 				codepointLen = 1;
 				break;
 			}
-			codepointLen++;
-			if (state == 0) {
+			++codepointLen;
+			if (state == UTF8_ACCEPT) {
 				break;
 			}
 		}
@@ -286,11 +285,11 @@ size_t charmap_ConvertNext(std::string_view &input, std::vector<int32_t> *output
 
 		// Warn if this character is not mapped but any others are
 		if (int firstChar = input[inputIdx]; charmap.nodes.size() > 1) {
-			warning(WARNING_UNMAPPED_CHAR_1, "Unmapped character %s\n", printChar(firstChar));
+			warning(WARNING_UNMAPPED_CHAR_1, "Unmapped character %s", printChar(firstChar));
 		} else if (charmap.name != DEFAULT_CHARMAP_NAME) {
 			warning(
 			    WARNING_UNMAPPED_CHAR_2,
-			    "Unmapped character %s not in " DEFAULT_CHARMAP_NAME " charmap\n",
+			    "Unmapped character %s not in " DEFAULT_CHARMAP_NAME " charmap",
 			    printChar(firstChar)
 			);
 		}
@@ -306,7 +305,7 @@ size_t charmap_ConvertNext(std::string_view &input, std::vector<int32_t> *output
 std::string charmap_Reverse(std::vector<int32_t> const &value, bool &unique) {
 	Charmap const &charmap = *currentCharmap;
 	std::string revMapping;
-	unique = charmap.forEachChar([&](size_t nodeIdx, std::string const &mapping) {
+	unique = forEachChar(charmap, [&](size_t nodeIdx, std::string const &mapping) {
 		if (charmap.nodes[nodeIdx].value == value) {
 			if (revMapping.empty()) {
 				revMapping = mapping;

src/asm/fixpoint.cpp

@@ -10,12 +10,6 @@
 	#define M_PI 3.14159265358979323846
 #endif
 
-uint8_t fixPrecision;
-
-uint8_t fix_Precision() {
-	return fixPrecision;
-}
-
 static double fix2double(int32_t i, int32_t q) {
 	return i / pow(2.0, q);
 }

src/asm/format.cpp

@@ -10,6 +10,7 @@
 #include <string.h>
 
 #include "asm/fixpoint.hpp"
+#include "asm/main.hpp" // options
 #include "asm/warning.hpp"
 
 void FormatSpec::useCharacter(int c) {
@@ -162,19 +163,19 @@ void FormatSpec::appendString(std::string &str, std::string const &value) const
 	}
 
 	if (sign) {
-		error("Formatting string with sign flag '%c'\n", sign);
+		error("Formatting string with sign flag '%c'", sign);
 	}
 	if (padZero) {
-		error("Formatting string with padding flag '0'\n");
+		error("Formatting string with padding flag '0'");
 	}
 	if (hasFrac) {
-		error("Formatting string with fractional width\n");
+		error("Formatting string with fractional width");
 	}
 	if (hasPrec) {
-		error("Formatting string with fractional precision\n");
+		error("Formatting string with fractional precision");
 	}
 	if (useType != 's') {
-		error("Formatting string as type '%c'\n", useType);
+		error("Formatting string as type '%c'", useType);
 	}
 
 	std::string useValue = exact ? escapeString(value) : value;
@@ -203,16 +204,16 @@ void FormatSpec::appendNumber(std::string &str, uint32_t value) const {
 
 	if (useType != 'X' && useType != 'x' && useType != 'b' && useType != 'o' && useType != 'f'
 	    && useExact) {
-		error("Formatting type '%c' with exact flag '#'\n", useType);
+		error("Formatting type '%c' with exact flag '#'", useType);
 	}
 	if (useType != 'f' && hasFrac) {
-		error("Formatting type '%c' with fractional width\n", useType);
+		error("Formatting type '%c' with fractional width", useType);
 	}
 	if (useType != 'f' && hasPrec) {
-		error("Formatting type '%c' with fractional precision\n", useType);
+		error("Formatting type '%c' with fractional precision", useType);
 	}
 	if (useType == 's') {
-		error("Formatting number as type 's'\n");
+		error("Formatting number as type 's'");
 	}
 
 	char signChar = sign; // 0 or ' ' or '+'
@@ -254,15 +255,15 @@ void FormatSpec::appendNumber(std::string &str, uint32_t value) const {
 		// Default fractional width (C++'s is 6 for "%f"; here 5 is enough for Q16.16)
 		size_t useFracWidth = hasFrac ? fracWidth : 5;
 		if (useFracWidth > 255) {
-			error("Fractional width %zu too long, limiting to 255\n", useFracWidth);
+			error("Fractional width %zu too long, limiting to 255", useFracWidth);
 			useFracWidth = 255;
 		}
 
-		size_t defaultPrec = fix_Precision();
+		size_t defaultPrec = options.fixPrecision;
 		size_t usePrec = hasPrec ? precision : defaultPrec;
 		if (usePrec < 1 || usePrec > 31) {
 			error(
-			    "Fixed-point constant precision %zu invalid, defaulting to %zu\n",
+			    "Fixed-point constant precision %zu invalid, defaulting to %zu",
 			    usePrec,
 			    defaultPrec
 			);

src/asm/fstack.cpp

@@ -3,6 +3,7 @@
 #include "asm/fstack.hpp"
 #include <sys/stat.h>
 
+#include <deque>
 #include <errno.h>
 #include <inttypes.h>
 #include <memory>
@@ -10,7 +11,7 @@
 #include <stdio.h>
 #include <stdlib.h>
 
-#include "error.hpp"
+#include "diagnostics.hpp"
 #include "helpers.hpp"
 #include "linkdefs.hpp"
 #include "platform.hpp" // S_ISDIR (stat macro)
@@ -41,23 +42,29 @@ struct Context {
 };
 
 static std::stack<Context> contextStack;
-size_t maxRecursionDepth;
 
 // The first include path for `fstk_FindFile` to try is none at all
-static std::vector<std::string> includePaths = {""};
+static std::vector<std::string> includePaths = {""}; // -I
+static std::deque<std::string> preIncludeNames;      // -P
+static bool failedOnMissingInclude = false;
 
-static std::string preIncludeName;
+std::string FileStackNode::reptChain() const {
+	std::string chain;
+	std::vector<uint32_t> const &nodeIters = iters();
+	for (uint32_t i = nodeIters.size(); i--;) {
+		chain.append("::REPT~");
+		chain.append(std::to_string(nodeIters[i]));
+	}
+	return chain;
+}
 
 std::string const &FileStackNode::dump(uint32_t curLineNo) const {
-	if (data.holds<std::vector<uint32_t>>()) {
+	if (std::holds_alternative<std::vector<uint32_t>>(data)) {
 		assume(parent); // REPT nodes use their parent's name
 		std::string const &lastName = parent->dump(lineNo);
 		fputs(" -> ", stderr);
 		fputs(lastName.c_str(), stderr);
-		std::vector<uint32_t> const &nodeIters = iters();
-		for (uint32_t i = nodeIters.size(); i--;) {
-			fprintf(stderr, "::REPT~%" PRIu32, nodeIters[i]);
-		}
+		fputs(reptChain().c_str(), stderr);
 		fprintf(stderr, "(%" PRIu32 ")", curLineNo);
 		return lastName;
 	} else {
@@ -72,12 +79,13 @@ std::string const &FileStackNode::dump(uint32_t curLineNo) const {
 	}
 }
 
-void fstk_DumpCurrent() {
+bool fstk_DumpCurrent() {
 	if (lexer_AtTopLevel()) {
-		fputs("at top level", stderr);
-		return;
+		return false;
 	}
+	assume(!contextStack.empty());
 	contextStack.top().fileInfo->dump(lexer_GetLineNo());
+	return true;
 }
 
 std::shared_ptr<FileStackNode> fstk_GetFileStack() {
@@ -114,16 +122,9 @@ void fstk_AddIncludePath(std::string const &path) {
 	}
 }
 
-void fstk_SetPreIncludeFile(std::string const &path) {
-	if (!preIncludeName.empty()) {
-		warnx("Overriding pre-included filename %s", preIncludeName.c_str());
-	}
-	preIncludeName = path;
-	// LCOV_EXCL_START
-	if (verbose) {
-		printf("Pre-included filename %s\n", preIncludeName.c_str());
-	}
-	// LCOV_EXCL_STOP
+void fstk_AddPreIncludeFile(std::string const &path) {
+	preIncludeNames.emplace_front(path);
+	verbosePrint("Pre-included filename %s\n", path.c_str()); // LCOV_EXCL_LINE
 }
 
 static bool isValidFilePath(std::string const &path) {
@@ -132,11 +133,9 @@ static bool isValidFilePath(std::string const &path) {
 }
 
 static void printDep(std::string const &path) {
-	if (dependFile) {
-		fprintf(dependFile, "%s: %s\n", targetFileName.c_str(), path.c_str());
-		if (generatePhonyDeps && isValidFilePath(path)) {
-			fprintf(dependFile, "%s:\n", path.c_str());
-		}
+	options.printDep(path);
+	if (options.dependFile && options.generatePhonyDeps && isValidFilePath(path)) {
+		fprintf(options.dependFile, "%s:\n", path.c_str());
 	}
 }
 
@@ -149,7 +148,7 @@ std::optional<std::string> fstk_FindFile(std::string const &path) {
 	}
 
 	errno = ENOENT;
-	if (generatedMissingIncludes) {
+	if (options.missingIncludeState != INC_ERROR) {
 		printDep(path);
 	}
 	return std::nullopt;
@@ -159,8 +158,8 @@ bool yywrap() {
 	uint32_t ifDepth = lexer_GetIFDepth();
 
 	if (ifDepth != 0) {
-		fatalerror(
-		    "Ended block with %" PRIu32 " unterminated IF construct%s\n",
+		fatal(
+		    "Ended block with %" PRIu32 " unterminated IF construct%s",
 		    ifDepth,
 		    ifDepth == 1 ? "" : "s"
 		);
@@ -188,11 +187,11 @@ bool yywrap() {
 
 			// This error message will refer to the current iteration
 			if (sym->type != SYM_VAR) {
-				fatalerror("Failed to update FOR symbol value\n");
+				fatal("Failed to update FOR symbol value");
 			}
 		}
 		// Advance to the next iteration
-		fileInfoIters.front()++;
+		++fileInfoIters.front();
 		// If this wasn't the last iteration, wrap instead of popping
 		if (fileInfoIters.front() <= context.nbReptIters) {
 			lexer_RestartRept(context.fileInfo->lineNo);
@@ -210,12 +209,12 @@ bool yywrap() {
 }
 
 static void checkRecursionDepth() {
-	if (contextStack.size() > maxRecursionDepth) {
-		fatalerror("Recursion limit (%zu) exceeded\n", maxRecursionDepth);
+	if (contextStack.size() > options.maxRecursionDepth) {
+		fatal("Recursion limit (%zu) exceeded", options.maxRecursionDepth);
 	}
 }
 
-static bool newFileContext(std::string const &filePath, bool updateStateNow) {
+static void newFileContext(std::string const &filePath, bool updateStateNow) {
 	checkRecursionDepth();
 
 	std::shared_ptr<std::string> uniqueIDStr = nullptr;
@@ -237,7 +236,7 @@ static bool newFileContext(std::string const &filePath, bool updateStateNow) {
 	    .macroArgs = macroArgs,
 	});
 
-	return context.lexerState.setFileAsNextState(filePath, updateStateNow);
+	context.lexerState.setFileAsNextState(filePath, updateStateNow);
 }
 
 static void newMacroContext(Symbol const &macro, std::shared_ptr<MacroArgs> macroArgs) {
@@ -253,11 +252,7 @@ static void newMacroContext(Symbol const &macro, std::shared_ptr<MacroArgs> macr
 		}
 	}
 	if (macro.src->type == NODE_REPT) {
-		std::vector<uint32_t> const &srcIters = macro.src->iters();
-		for (uint32_t i = srcIters.size(); i--;) {
-			fileInfoName.append("::REPT~");
-			fileInfoName.append(std::to_string(srcIters[i]));
-		}
+		fileInfoName.append(macro.src->reptChain());
 	}
 	fileInfoName.append("::");
 	fileInfoName.append(macro.name);
@@ -305,26 +300,34 @@ static Context &newReptContext(int32_t reptLineNo, ContentSpan const &span, uint
 	return context;
 }
 
-void fstk_RunInclude(std::string const &path, bool preInclude) {
-	std::optional<std::string> fullPath = fstk_FindFile(path);
-
-	if (!fullPath) {
-		if (generatedMissingIncludes && !preInclude) {
-			// LCOV_EXCL_START
-			if (verbose) {
-				printf("Aborting (-MG) on INCLUDE file '%s' (%s)\n", path.c_str(), strerror(errno));
-			}
-			// LCOV_EXCL_STOP
-			failedOnMissingInclude = true;
+bool fstk_FileError(std::string const &path, char const *functionName) {
+	if (options.missingIncludeState == INC_ERROR) {
+		error("Error opening %s file '%s': %s", functionName, path.c_str(), strerror(errno));
 	} else {
-			error("Unable to open included file '%s': %s\n", path.c_str(), strerror(errno));
+		failedOnMissingInclude = true;
+		// LCOV_EXCL_START
+		if (options.missingIncludeState == GEN_EXIT) {
+			verbosePrint(
+			    "Aborting (-MG) on %s file '%s' (%s)\n", functionName, path.c_str(), strerror(errno)
+			);
+			return true;
 		}
-		return;
+		assume(options.missingIncludeState == GEN_CONTINUE);
+		// LCOV_EXCL_STOP
 	}
+	return false;
+}
 
-	if (!newFileContext(*fullPath, false)) {
-		fatalerror("Failed to set up lexer for file include\n"); // LCOV_EXCL_LINE
+bool fstk_FailedOnMissingInclude() {
+	return failedOnMissingInclude;
+}
+
+bool fstk_RunInclude(std::string const &path) {
+	if (std::optional<std::string> fullPath = fstk_FindFile(path); fullPath) {
+		newFileContext(*fullPath, false);
+		return false;
 	}
+	return fstk_FileError(path, "INCLUDE");
 }
 
 void fstk_RunMacro(std::string const &macroName, std::shared_ptr<MacroArgs> macroArgs) {
@@ -332,14 +335,14 @@ void fstk_RunMacro(std::string const &macroName, std::shared_ptr<MacroArgs> macr
 
 	if (!macro) {
 		if (sym_IsPurgedExact(macroName)) {
-			error("Macro \"%s\" not defined; it was purged\n", macroName.c_str());
+			error("Macro \"%s\" not defined; it was purged", macroName.c_str());
 		} else {
-			error("Macro \"%s\" not defined\n", macroName.c_str());
+			error("Macro \"%s\" not defined", macroName.c_str());
 		}
 		return;
 	}
 	if (macro->type != SYM_MACRO) {
-		error("\"%s\" is not a macro\n", macroName.c_str());
+		error("\"%s\" is not a macro", macroName.c_str());
 		return;
 	}
 
@@ -372,13 +375,11 @@ void fstk_RunFor(
 	} else if (step < 0 && stop < start) {
 		count = (static_cast<int64_t>(start) - stop - 1) / -static_cast<int64_t>(step) + 1;
 	} else if (step == 0) {
-		error("FOR cannot have a step value of 0\n");
+		error("FOR cannot have a step value of 0");
 	}
 
 	if ((step > 0 && start > stop) || (step < 0 && start < stop)) {
-		warning(
-		    WARNING_BACKWARDS_FOR, "FOR goes backwards from %d to %d by %d\n", start, stop, step
-		);
+		warning(WARNING_BACKWARDS_FOR, "FOR goes backwards from %d to %d by %d", start, stop, step);
 	}
 
 	if (count == 0) {
@@ -394,7 +395,7 @@ void fstk_RunFor(
 
 bool fstk_Break() {
 	if (contextStack.top().fileInfo->type != NODE_REPT) {
-		error("BREAK can only be used inside a REPT/FOR block\n");
+		error("BREAK can only be used inside a REPT/FOR block");
 		return false;
 	}
 
@@ -404,19 +405,19 @@ bool fstk_Break() {
 
 void fstk_NewRecursionDepth(size_t newDepth) {
 	if (contextStack.size() > newDepth + 1) {
-		fatalerror("Recursion limit (%zu) exceeded\n", newDepth);
+		fatal("Recursion limit (%zu) exceeded", newDepth);
 	}
-	maxRecursionDepth = newDepth;
+	options.maxRecursionDepth = newDepth;
 }
 
-void fstk_Init(std::string const &mainPath, size_t maxDepth) {
-	if (!newFileContext(mainPath, true)) {
-		fatalerror("Failed to open main file\n");
+void fstk_Init(std::string const &mainPath) {
+	newFileContext(mainPath, true);
+
+	for (std::string const &name : preIncludeNames) {
+		if (std::optional<std::string> fullPath = fstk_FindFile(name); fullPath) {
+			newFileContext(*fullPath, false);
+		} else {
+			error("Error reading pre-included file '%s': %s", name.c_str(), strerror(errno));
 		}
-
-	maxRecursionDepth = maxDepth;
-
-	if (!preIncludeName.empty()) {
-		fstk_RunInclude(preIncludeName, true);
 	}
 }

src/asm/macro.cpp

@@ -29,15 +29,15 @@ std::shared_ptr<std::string> MacroArgs::getAllArgs() const {
 
 	size_t len = 0;
 
-	for (uint32_t i = shift; i < nbArgs; i++) {
+	for (uint32_t i = shift; i < nbArgs; ++i) {
 		len += args[i]->length() + 1; // 1 for comma
 	}
 
 	auto str = std::make_shared<std::string>();
 	str->reserve(len + 1); // 1 for comma
 
-	for (uint32_t i = shift; i < nbArgs; i++) {
-		auto const &arg = args[i];
+	for (uint32_t i = shift; i < nbArgs; ++i) {
+		std::shared_ptr<std::string> const &arg = args[i];
 
 		str->append(*arg);
 
@@ -52,7 +52,7 @@ std::shared_ptr<std::string> MacroArgs::getAllArgs() const {
 
 void MacroArgs::appendArg(std::shared_ptr<std::string> arg) {
 	if (arg->empty()) {
-		warning(WARNING_EMPTY_MACRO_ARG, "Empty macro argument\n");
+		warning(WARNING_EMPTY_MACRO_ARG, "Empty macro argument");
 	}
 	args.push_back(arg);
 }
@@ -60,10 +60,10 @@ void MacroArgs::appendArg(std::shared_ptr<std::string> arg) {
 void MacroArgs::shiftArgs(int32_t count) {
 	if (size_t nbArgs = args.size();
 	    count > 0 && (static_cast<uint32_t>(count) > nbArgs || shift > nbArgs - count)) {
-		warning(WARNING_MACRO_SHIFT, "Cannot shift macro arguments past their end\n");
+		warning(WARNING_MACRO_SHIFT, "Cannot shift macro arguments past their end");
 		shift = nbArgs;
 	} else if (count < 0 && shift < static_cast<uint32_t>(-count)) {
-		warning(WARNING_MACRO_SHIFT, "Cannot shift macro arguments past their beginning\n");
+		warning(WARNING_MACRO_SHIFT, "Cannot shift macro arguments past their beginning");
 		shift = 0;
 	} else {
 		shift += count;

src/asm/main.cpp

@@ -3,16 +3,20 @@
 #include "asm/main.hpp"
 
 #include <algorithm>
+#include <errno.h>
 #include <limits.h>
 #include <memory>
+#include <stdarg.h>
 #include <stdlib.h>
 #include <string.h>
 #include <time.h>
 
-#include "error.hpp"
+#include "diagnostics.hpp"
 #include "extern/getopt.hpp"
 #include "helpers.hpp"
-#include "parser.hpp"
+#include "parser.hpp" // Generated from parser.y
+#include "usage.hpp"
+#include "util.hpp" // UpperMap
 #include "version.hpp"
 
 #include "asm/charmap.hpp"
@@ -22,13 +26,7 @@
 #include "asm/symbol.hpp"
 #include "asm/warning.hpp"
 
-FILE *dependFile = nullptr;            // -M
-bool generatedMissingIncludes = false; // -MG
-bool generatePhonyDeps = false;        // -MP
-std::string targetFileName;            // -MQ, -MT
-bool failedOnMissingInclude = false;
-bool verbose = false; // -v
-bool warnings = true; // -w
+Options options;
 
 // Escapes Make-special chars from a string
 static std::string make_escape(std::string &str) {
@@ -66,14 +64,14 @@ static option const longopts[] = {
     {"define",          required_argument, nullptr,  'D'},
     {"export-all",      no_argument,       nullptr,  'E'},
     {"gfx-chars",       required_argument, nullptr,  'g'},
+    {"help",            no_argument,       nullptr,  'h'},
     {"include",         required_argument, nullptr,  'I'},
     {"dependfile",      required_argument, nullptr,  'M'},
+    {"MC",              no_argument,       &depType, 'C'},
     {"MG",              no_argument,       &depType, 'G'},
-    {"help",            no_argument,       nullptr,  'h'},
     {"MP",              no_argument,       &depType, 'P'},
-    {"MT",              required_argument, &depType, 'T'},
-    {"warning",         required_argument, nullptr,  'W'},
     {"MQ",              required_argument, &depType, 'Q'},
+    {"MT",              required_argument, &depType, 'T'},
     {"output",          required_argument, nullptr,  'o'},
     {"preinclude",      required_argument, nullptr,  'P'},
     {"pad-value",       required_argument, nullptr,  'p'},
@@ -87,11 +85,10 @@ static option const longopts[] = {
     {nullptr,           no_argument,       nullptr,  0  }
 };
 
-// LCOV_EXCL_START
-static void printUsage() {
-	fputs(
+// clang-format off: long string literal
+static Usage usage(
     "Usage: rgbasm [-EhVvw] [-b chars] [-D name[=value]] [-g chars] [-I path]\n"
-	    "              [-M depend_file] [-MG] [-MP] [-MT target_file] [-MQ target_file]\n"
+    "              [-M depend_file] [-MC] [-MG] [-MP] [-MT target_file] [-MQ target_file]\n"
     "              [-o out_file] [-P include_file] [-p pad_value] [-Q precision]\n"
     "              [-r depth] [-s features:state_file] [-W warning] [-X max_errors]\n"
     "              <file>\n"
@@ -104,44 +101,73 @@ static void printUsage() {
     "    -V, --version                  print RGBASM version and exit\n"
     "    -W, --warning <warning>        enable or disable warnings\n"
     "\n"
-	    "For help, use `man rgbasm' or go to https://rgbds.gbdev.io/docs/\n",
-	    stderr
-	);
+    "For help, use `man rgbasm' or go to https://rgbds.gbdev.io/docs/\n"
+);
+// clang-format on
+
+// Parse a comma-separated string of '-s/--state' features
+static std::vector<StateFeature> parseStateFeatures(char *str) {
+	std::vector<StateFeature> features;
+	for (char *feature = str; feature;) {
+		// Split "<feature>,<rest>" so `feature` is "<feature>" and `next` is "<rest>"
+		char *next = strchr(feature, ',');
+		if (next) {
+			*next++ = '\0';
+		}
+		// Trim whitespace from the beginning of `feature`...
+		feature += strspn(feature, " \t");
+		// ...and from the end
+		if (char *end = strpbrk(feature, " \t"); end) {
+			*end = '\0';
+		}
+		// A feature must be specified
+		if (*feature == '\0') {
+			fatal("Empty feature for option 's'");
+		}
+		// Parse the `feature` and update the `features` list
+		static UpperMap<StateFeature> const featureNames{
+		    {"EQU",   STATE_EQU  },
+		    {"VAR",   STATE_VAR  },
+		    {"EQUS",  STATE_EQUS },
+		    {"CHAR",  STATE_CHAR },
+		    {"MACRO", STATE_MACRO},
+		};
+		if (!strcasecmp(feature, "all")) {
+			if (!features.empty()) {
+				warnx("Redundant feature before \"%s\" for option 's'", feature);
+			}
+			features.assign({STATE_EQU, STATE_VAR, STATE_EQUS, STATE_CHAR, STATE_MACRO});
+		} else if (auto search = featureNames.find(feature); search == featureNames.end()) {
+			fatal("Invalid feature for option 's': \"%s\"", feature);
+		} else if (StateFeature value = search->second;
+		           std::find(RANGE(features), value) != features.end()) {
+			warnx("Ignoring duplicate feature for option 's': \"%s\"", feature);
+		} else {
+			features.push_back(value);
+		}
+		feature = next;
+	}
+	return features;
 }
-// LCOV_EXCL_STOP
 
 int main(int argc, char *argv[]) {
-	time_t now = time(nullptr);
 	// Support SOURCE_DATE_EPOCH for reproducible builds
 	// https://reproducible-builds.org/docs/source-date-epoch/
+	time_t now = time(nullptr);
 	if (char const *sourceDateEpoch = getenv("SOURCE_DATE_EPOCH"); sourceDateEpoch) {
 		now = static_cast<time_t>(strtoul(sourceDateEpoch, nullptr, 0));
 	}
-
-	Defer closeDependFile{[&] {
-		if (dependFile) {
-			fclose(dependFile);
-		}
-	}};
-
-	// Perform some init for below
 	sym_Init(now);
 
-	// Set defaults
-	opt_B("01");
-	opt_G("0123");
-	opt_P(0);
-	opt_Q(16);
-	sym_SetExportAll(false);
-	uint32_t maxDepth = 64;
-	char const *dependFileName = nullptr;
-	std::unordered_map<std::string, std::vector<StateFeature>> stateFileSpecs;
-	std::string newTarget;
-	// Maximum of 100 errors only applies if rgbasm is printing errors to a terminal.
+	// Maximum of 100 errors only applies if rgbasm is printing errors to a terminal
 	if (isatty(STDERR_FILENO)) {
-		maxErrors = 100;
+		options.maxErrors = 100;
 	}
 
+	// Local options
+	char const *dependFileName = nullptr;                                      // -M
+	std::unordered_map<std::string, std::vector<StateFeature>> stateFileSpecs; // -s
+
 	for (int ch; (ch = musl_getopt_long_only(argc, argv, optstring, longopts, nullptr)) != -1;) {
 		switch (ch) {
 			char *endptr;
@@ -150,7 +176,7 @@ int main(int argc, char *argv[]) {
 			if (strlen(musl_optarg) == 2) {
 				opt_B(musl_optarg);
 			} else {
-				errx("Must specify exactly 2 characters for option 'b'");
+				fatal("Must specify exactly 2 characters for option 'b'");
 			}
 			break;
 
@@ -173,42 +199,45 @@ int main(int argc, char *argv[]) {
 			if (strlen(musl_optarg) == 4) {
 				opt_G(musl_optarg);
 			} else {
-				errx("Must specify exactly 4 characters for option 'g'");
+				fatal("Must specify exactly 4 characters for option 'g'");
 			}
 			break;
 
 		case 'h':
-			// LCOV_EXCL_START
-			printUsage();
-			exit(0);
-			// LCOV_EXCL_STOP
+			usage.printAndExit(0); // LCOV_EXCL_LINE
 
 		case 'I':
 			fstk_AddIncludePath(musl_optarg);
 			break;
 
 		case 'M':
-			if (dependFile) {
+			if (options.dependFile) {
 				warnx("Overriding dependfile %s", dependFileName);
 			}
 			if (strcmp("-", musl_optarg)) {
-				dependFile = fopen(musl_optarg, "w");
+				options.dependFile = fopen(musl_optarg, "w");
 				dependFileName = musl_optarg;
 			} else {
-				dependFile = stdout;
+				options.dependFile = stdout;
 				dependFileName = "<stdout>";
 			}
-			if (dependFile == nullptr) {
-				err("Failed to open dependfile \"%s\"", dependFileName); // LCOV_EXCL_LINE
+			if (options.dependFile == nullptr) {
+				// LCOV_EXCL_START
+				fatal("Failed to open dependfile \"%s\": %s", dependFileName, strerror(errno));
+				// LCOV_EXCL_STOP
 			}
 			break;
 
 		case 'o':
-			out_SetFileName(musl_optarg);
+			if (!options.objectFileName.empty()) {
+				warnx("Overriding output filename %s", options.objectFileName.c_str());
+			}
+			options.objectFileName = musl_optarg;
+			verbosePrint("Output filename %s\n", options.objectFileName.c_str()); // LCOV_EXCL_LINE
 			break;
 
 		case 'P':
-			fstk_SetPreIncludeFile(musl_optarg);
+			fstk_AddPreIncludeFile(musl_optarg);
 			break;
 
 			unsigned long padByte;
@@ -216,41 +245,40 @@ int main(int argc, char *argv[]) {
 			padByte = strtoul(musl_optarg, &endptr, 0);
 
 			if (musl_optarg[0] == '\0' || *endptr != '\0') {
-				errx("Invalid argument for option 'p'");
+				fatal("Invalid argument for option 'p'");
 			}
 
 			if (padByte > 0xFF) {
-				errx("Argument for option 'p' must be between 0 and 0xFF");
+				fatal("Argument for option 'p' must be between 0 and 0xFF");
 			}
 
 			opt_P(padByte);
 			break;
 
-			unsigned long precision;
-			char const *precisionArg;
-		case 'Q':
-			precisionArg = musl_optarg;
+		case 'Q': {
+			char const *precisionArg = musl_optarg;
 			if (precisionArg[0] == '.') {
-				precisionArg++;
+				++precisionArg;
 			}
-			precision = strtoul(precisionArg, &endptr, 0);
+			unsigned long precision = strtoul(precisionArg, &endptr, 0);
 
 			if (musl_optarg[0] == '\0' || *endptr != '\0') {
-				errx("Invalid argument for option 'Q'");
+				fatal("Invalid argument for option 'Q'");
 			}
 
 			if (precision < 1 || precision > 31) {
-				errx("Argument for option 'Q' must be between 1 and 31");
+				fatal("Argument for option 'Q' must be between 1 and 31");
 			}
 
 			opt_Q(precision);
 			break;
+		}
 
 		case 'r':
-			maxDepth = strtoul(musl_optarg, &endptr, 0);
+			options.maxRecursionDepth = strtoul(musl_optarg, &endptr, 0);
 
 			if (musl_optarg[0] == '\0' || *endptr != '\0') {
-				errx("Invalid argument for option 'r'");
+				fatal("Invalid argument for option 'r'");
 			}
 			break;
 
@@ -258,59 +286,16 @@ int main(int argc, char *argv[]) {
 			// Split "<features>:<name>" so `musl_optarg` is "<features>" and `name` is "<name>"
 			char *name = strchr(musl_optarg, ':');
 			if (!name) {
-				errx("Invalid argument for option 's'");
+				fatal("Invalid argument for option 's'");
 			}
 			*name++ = '\0';
 
-			std::vector<StateFeature> features;
-			for (char *feature = musl_optarg; feature;) {
-				// Split "<feature>,<rest>" so `feature` is "<feature>" and `next` is "<rest>"
-				char *next = strchr(feature, ',');
-				if (next) {
-					*next++ = '\0';
-				}
-				// Trim whitespace from the beginning of `feature`...
-				feature += strspn(feature, " \t");
-				// ...and from the end
-				if (char *end = strpbrk(feature, " \t"); end) {
-					*end = '\0';
-				}
-				// A feature must be specified
-				if (*feature == '\0') {
-					errx("Empty feature for option 's'");
-				}
-				// Parse the `feature` and update the `features` list
-				if (!strcasecmp(feature, "all")) {
-					if (!features.empty()) {
-						warnx("Redundant feature before \"%s\" for option 's'", feature);
-					}
-					features.assign({STATE_EQU, STATE_VAR, STATE_EQUS, STATE_CHAR, STATE_MACRO});
-				} else {
-					StateFeature value = !strcasecmp(feature, "equ")     ? STATE_EQU
-					                     : !strcasecmp(feature, "var")   ? STATE_VAR
-					                     : !strcasecmp(feature, "equs")  ? STATE_EQUS
-					                     : !strcasecmp(feature, "char")  ? STATE_CHAR
-					                     : !strcasecmp(feature, "macro") ? STATE_MACRO
-					                                                     : NB_STATE_FEATURES;
-					if (value == NB_STATE_FEATURES) {
-						errx("Invalid feature for option 's': \"%s\"", feature);
-					} else if (std::find(RANGE(features), value) != features.end()) {
-						warnx("Ignoring duplicate feature for option 's': \"%s\"", feature);
-					} else {
-						features.push_back(value);
-					}
-				}
-				feature = next;
-			}
+			std::vector<StateFeature> features = parseStateFeatures(musl_optarg);
 
 			if (stateFileSpecs.find(name) != stateFileSpecs.end()) {
 				warnx("Overriding state filename %s", name);
 			}
-			// LCOV_EXCL_START
-			if (verbose) {
-				printf("State filename %s\n", name);
-			}
-			// LCOV_EXCL_STOP
+			verbosePrint("State filename %s\n", name); // LCOV_EXCL_LINE
 			stateFileSpecs.emplace(name, std::move(features));
 			break;
 		}
@@ -321,7 +306,7 @@ int main(int argc, char *argv[]) {
 
 		case 'v':
 			// LCOV_EXCL_START
-			verbose = true;
+			options.verbose = true;
 			break;
 			// LCOV_EXCL_STOP
 
@@ -330,100 +315,93 @@ int main(int argc, char *argv[]) {
 			break;
 
 		case 'w':
-			warnings = false;
+			warnings.state.warningsEnabled = false;
 			break;
 
-			unsigned long maxValue;
-		case 'X':
-			maxValue = strtoul(musl_optarg, &endptr, 0);
+		case 'X': {
+			uint64_t maxErrors = strtoul(musl_optarg, &endptr, 0);
 
 			if (musl_optarg[0] == '\0' || *endptr != '\0') {
-				errx("Invalid argument for option 'X'");
+				fatal("Invalid argument for option 'X'");
 			}
 
-			if (maxValue > UINT_MAX) {
-				errx("Argument for option 'X' must be between 0 and %u", UINT_MAX);
+			if (maxErrors > UINT64_MAX) {
+				fatal("Argument for option 'X' must be between 0 and %" PRIu64, UINT64_MAX);
 			}
 
-			maxErrors = maxValue;
+			options.maxErrors = maxErrors;
 			break;
+		}
 
 		// Long-only options
 		case 0:
 			switch (depType) {
+			case 'C':
+				options.missingIncludeState = GEN_CONTINUE;
+				break;
+
 			case 'G':
-				generatedMissingIncludes = true;
+				options.missingIncludeState = GEN_EXIT;
 				break;
 
 			case 'P':
-				generatePhonyDeps = true;
+				options.generatePhonyDeps = true;
 				break;
 
 			case 'Q':
-			case 'T':
-				newTarget = musl_optarg;
+			case 'T': {
+				std::string newTarget = musl_optarg;
 				if (depType == 'Q') {
 					newTarget = make_escape(newTarget);
 				}
-				if (!targetFileName.empty()) {
-					targetFileName += ' ';
+				if (!options.targetFileName.empty()) {
+					options.targetFileName += ' ';
 				}
-				targetFileName += newTarget;
+				options.targetFileName += newTarget;
 				break;
 			}
+			}
 			break;
 
 		// Unrecognized options
 		default:
-			// LCOV_EXCL_START
-			printUsage();
-			exit(1);
-			// LCOV_EXCL_STOP
+			usage.printAndExit(1); // LCOV_EXCL_LINE
 		}
 	}
 
-	if (targetFileName.empty() && !objectFileName.empty()) {
-		targetFileName = objectFileName;
+	if (options.targetFileName.empty() && !options.objectFileName.empty()) {
+		options.targetFileName = options.objectFileName;
 	}
 
 	if (argc == musl_optind) {
-		fputs(
-		    "FATAL: Please specify an input file (pass `-` to read from standard input)\n", stderr
-		);
-		printUsage();
-		exit(1);
+		usage.printAndExit("Please specify an input file (pass `-` to read from standard input)");
 	} else if (argc != musl_optind + 1) {
-		fputs("FATAL: More than one input file specified\n", stderr);
-		printUsage();
-		exit(1);
+		usage.printAndExit("More than one input file specified");
 	}
 
 	std::string mainFileName = argv[musl_optind];
 
-	if (verbose) {
-		printf("Assembling %s\n", mainFileName.c_str()); // LCOV_EXCL_LINE
-	}
+	verbosePrint("Assembling %s\n", mainFileName.c_str()); // LCOV_EXCL_LINE
 
-	if (dependFile) {
-		if (targetFileName.empty()) {
-			errx("Dependency files can only be created if a target file is specified with either "
+	if (options.dependFile && options.targetFileName.empty()) {
+		fatal("Dependency files can only be created if a target file is specified with either "
 		      "-o, -MQ or -MT");
 	}
-
-		fprintf(dependFile, "%s: %s\n", targetFileName.c_str(), mainFileName.c_str());
-	}
+	options.printDep(mainFileName);
 
 	charmap_New(DEFAULT_CHARMAP_NAME, nullptr);
 
 	// Init lexer and file stack, providing file info
-	fstk_Init(mainFileName, maxDepth);
+	fstk_Init(mainFileName);
 
 	// Perform parse (`yy::parser` is auto-generated from `parser.y`)
-	if (yy::parser parser; parser.parse() != 0 && nbErrors == 0) {
-		nbErrors = 1;
+	if (yy::parser parser; parser.parse() != 0) {
+		if (warnings.nbErrors == 0) {
+			warnings.nbErrors = 1;
+		}
 	}
 
-	if (!failedOnMissingInclude) {
+	if (!fstk_FailedOnMissingInclude()) {
 		sect_CheckUnionClosed();
 		sect_CheckLoadClosed();
 		sect_CheckSizes();
@@ -433,12 +411,10 @@ int main(int argc, char *argv[]) {
 		sect_CheckStack();
 	}
 
-	if (nbErrors != 0) {
-		errx("Assembly aborted (%u error%s)!", nbErrors, nbErrors == 1 ? "" : "s");
-	}
+	requireZeroErrors();
 
 	// If parse aborted due to missing an include, and `-MG` was given, exit normally
-	if (failedOnMissingInclude) {
+	if (fstk_FailedOnMissingInclude()) {
 		return 0;
 	}
 

src/asm/opt.cpp

@@ -7,9 +7,12 @@
 #include <stdlib.h>
 #include <string.h>
 
+#include "helpers.hpp" // assume
+
 #include "asm/fixpoint.hpp"
 #include "asm/fstack.hpp"
 #include "asm/lexer.hpp"
+#include "asm/main.hpp" // options
 #include "asm/section.hpp"
 #include "asm/warning.hpp"
 
@@ -17,37 +20,38 @@ struct OptStackEntry {
 	char binDigits[2];
 	char gfxDigits[4];
 	uint8_t fixPrecision;
-	uint8_t fillByte;
-	bool warningsAreErrors;
+	uint8_t padByte;
 	size_t maxRecursionDepth;
-	Diagnostics warningStates;
+	DiagnosticsState<WarningID> warningStates;
 };
 
 static std::stack<OptStackEntry> stack;
 
-void opt_B(char const chars[2]) {
-	lexer_SetBinDigits(chars);
+void opt_B(char const binDigits[2]) {
+	lexer_SetBinDigits(binDigits);
 }
 
-void opt_G(char const chars[4]) {
-	lexer_SetGfxDigits(chars);
+void opt_G(char const gfxDigits[4]) {
+	lexer_SetGfxDigits(gfxDigits);
 }
 
 void opt_P(uint8_t padByte) {
-	fillByte = padByte;
+	options.padByte = padByte;
 }
 
-void opt_Q(uint8_t precision) {
-	fixPrecision = precision;
+void opt_Q(uint8_t fixPrecision) {
+	options.fixPrecision = fixPrecision;
 }
 
-void opt_R(size_t newDepth) {
-	fstk_NewRecursionDepth(newDepth);
+void opt_R(size_t maxRecursionDepth) {
+	fstk_NewRecursionDepth(maxRecursionDepth);
 	lexer_CheckRecursionDepth();
 }
 
 void opt_W(char const *flag) {
-	processWarningFlag(flag);
+	if (warnings.processWarningFlag(flag) == "numeric-string") {
+		warning(WARNING_OBSOLETE, "Warning flag \"numeric-string\" is deprecated");
+	}
 }
 
 void opt_Parse(char const *s) {
@@ -56,7 +60,7 @@ void opt_Parse(char const *s) {
 		if (strlen(&s[1]) == 2) {
 			opt_B(&s[1]);
 		} else {
-			error("Must specify exactly 2 characters for option 'b'\n");
+			error("Must specify exactly 2 characters for option 'b'");
 		}
 		break;
 
@@ -64,7 +68,7 @@ void opt_Parse(char const *s) {
 		if (strlen(&s[1]) == 4) {
 			opt_G(&s[1]);
 		} else {
-			error("Must specify exactly 4 characters for option 'g'\n");
+			error("Must specify exactly 4 characters for option 'g'");
 		}
 		break;
 
@@ -75,14 +79,14 @@ void opt_Parse(char const *s) {
 
 			result = sscanf(&s[1], "%x", &padByte);
 			if (result != 1) {
-				error("Invalid argument for option 'p'\n");
-			} else if (padByte > 0xFF) {
-				error("Argument for option 'p' must be between 0 and 0xFF\n");
+				error("Invalid argument for option 'p'");
 			} else {
+				// Two characters cannot be scanned as a hex number greater than 0xFF
+				assume(padByte <= 0xFF);
 				opt_P(padByte);
 			}
 		} else {
-			error("Invalid argument for option 'p'\n");
+			error("Invalid argument for option 'p'");
 		}
 		break;
 
@@ -90,22 +94,22 @@ void opt_Parse(char const *s) {
 	case 'Q':
 		precisionArg = &s[1];
 		if (precisionArg[0] == '.') {
-			precisionArg++;
+			++precisionArg;
 		}
 		if (strlen(precisionArg) <= 2) {
 			int result;
-			unsigned int precision;
+			unsigned int fixPrecision;
 
-			result = sscanf(precisionArg, "%u", &precision);
+			result = sscanf(precisionArg, "%u", &fixPrecision);
 			if (result != 1) {
-				error("Invalid argument for option 'Q'\n");
-			} else if (precision < 1 || precision > 31) {
-				error("Argument for option 'Q' must be between 1 and 31\n");
+				error("Invalid argument for option 'Q'");
+			} else if (fixPrecision < 1 || fixPrecision > 31) {
+				error("Argument for option 'Q' must be between 1 and 31");
 			} else {
-				opt_Q(precision);
+				opt_Q(fixPrecision);
 			}
 		} else {
-			error("Invalid argument for option 'Q'\n");
+			error("Invalid argument for option 'Q'");
 		}
 		break;
 
@@ -116,19 +120,19 @@ void opt_Parse(char const *s) {
 		}
 
 		if (s[0] == '\0') {
-			error("Missing argument to option 'r'\n");
+			error("Missing argument to option 'r'");
 			break;
 		}
 
 		char *endptr;
-		unsigned long newDepth = strtoul(s, &endptr, 10);
+		unsigned long maxRecursionDepth = strtoul(s, &endptr, 10);
 
 		if (*endptr != '\0') {
-			error("Invalid argument to option 'r' (\"%s\")\n", s);
+			error("Invalid argument to option 'r' (\"%s\")", s);
 		} else if (errno == ERANGE) {
-			error("Argument to 'r' is out of range (\"%s\")\n", s);
+			error("Argument to 'r' is out of range (\"%s\")", s);
 		} else {
-			opt_R(newDepth);
+			opt_R(maxRecursionDepth);
 		}
 		break;
 	}
@@ -137,12 +141,12 @@ void opt_Parse(char const *s) {
 		if (strlen(&s[1]) > 0) {
 			opt_W(&s[1]);
 		} else {
-			error("Must specify an argument for option 'W'\n");
+			error("Must specify an argument for option 'W'");
 		}
 		break;
 
 	default:
-		error("Unknown option '%c'\n", s[0]);
+		error("Unknown option '%c'", s[0]);
 		break;
 	}
 }
@@ -150,26 +154,19 @@ void opt_Parse(char const *s) {
 void opt_Push() {
 	OptStackEntry entry;
 
-	// Both of these are pulled from lexer.hpp
-	memcpy(entry.binDigits, binDigits, std::size(binDigits));
-	memcpy(entry.gfxDigits, gfxDigits, std::size(gfxDigits));
-
-	entry.fixPrecision = fixPrecision; // Pulled from fixpoint.hpp
-
-	entry.fillByte = fillByte; // Pulled from section.hpp
-
-	// Both of these pulled from warning.hpp
-	entry.warningsAreErrors = warningsAreErrors;
-	entry.warningStates = warningStates;
-
-	entry.maxRecursionDepth = maxRecursionDepth; // Pulled from fstack.h
+	memcpy(entry.binDigits, options.binDigits, std::size(options.binDigits));
+	memcpy(entry.gfxDigits, options.gfxDigits, std::size(options.gfxDigits));
+	entry.padByte = options.padByte;
+	entry.fixPrecision = options.fixPrecision;
+	entry.maxRecursionDepth = options.maxRecursionDepth;
+	entry.warningStates = warnings.state;
 
 	stack.push(entry);
 }
 
 void opt_Pop() {
 	if (stack.empty()) {
-		error("No entries in the option stack\n");
+		error("No entries in the option stack");
 		return;
 	}
 
@@ -178,17 +175,16 @@ void opt_Pop() {
 
 	opt_B(entry.binDigits);
 	opt_G(entry.gfxDigits);
-	opt_P(entry.fillByte);
+	opt_P(entry.padByte);
 	opt_Q(entry.fixPrecision);
 	opt_R(entry.maxRecursionDepth);
 
-	// opt_W does not apply a whole warning state; it processes one flag string
-	warningsAreErrors = entry.warningsAreErrors;
-	warningStates = entry.warningStates;
+	// `opt_W` does not apply a whole warning state; it processes one flag string
+	warnings.state = entry.warningStates;
 }
 
 void opt_CheckStack() {
 	if (!stack.empty()) {
-		warning(WARNING_UNMATCHED_DIRECTIVE, "`PUSHO` without corresponding `POPO`\n");
+		warning(WARNING_UNMATCHED_DIRECTIVE, "`PUSHO` without corresponding `POPO`");
 	}
 }

src/asm/output.cpp

@@ -4,13 +4,14 @@
 
 #include <algorithm>
 #include <deque>
+#include <errno.h>
 #include <inttypes.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <string>
 
-#include "error.hpp"
+#include "diagnostics.hpp"
 #include "helpers.hpp" // assume, Defer
 #include "platform.hpp"
 
@@ -29,8 +30,6 @@ struct Assertion {
 	std::string message;
 };
 
-std::string objectFileName;
-
 // List of symbols to put in the object file
 static std::vector<Symbol *> objectSymbols;
 
@@ -61,27 +60,13 @@ void out_RegisterNode(std::shared_ptr<FileStackNode> node) {
 	}
 }
 
-// Return a section's ID, or UINT32_MAX if the section does not exist
-static uint32_t getSectIDIfAny(Section *sect) {
-	if (!sect) {
-		return UINT32_MAX;
-	}
-
-	if (auto search = sectionMap.find(sect->name); search != sectionMap.end()) {
-		return static_cast<uint32_t>(search->second);
-	}
-
-	// Every section that exists should be in `sectionMap`
-	fatalerror("Unknown section '%s'\n", sect->name.c_str()); // LCOV_EXCL_LINE
-}
-
 static void writePatch(Patch const &patch, FILE *file) {
 	assume(patch.src->ID != UINT32_MAX);
 
 	putLong(patch.src->ID, file);
 	putLong(patch.lineNo, file);
 	putLong(patch.offset, file);
-	putLong(getSectIDIfAny(patch.pcSection), file);
+	putLong(patch.pcSection ? patch.pcSection->getID() : UINT32_MAX, file);
 	putLong(patch.pcOffset, file);
 	putc(patch.type, file);
 	putLong(patch.rpn.size(), file);
@@ -125,10 +110,12 @@ static void writeSymbol(Symbol const &sym, FILE *file) {
 	} else {
 		assume(sym.src->ID != UINT32_MAX);
 
+		Section *symSection = sym.getSection();
+
 		putc(sym.isExported ? SYMTYPE_EXPORT : SYMTYPE_LOCAL, file);
 		putLong(sym.src->ID, file);
 		putLong(sym.fileLine, file);
-		putLong(getSectIDIfAny(sym.getSection()), file);
+		putLong(symSection ? symSection->getID() : UINT32_MAX, file);
 		putLong(sym.getOutputValue(), file);
 	}
 }
@@ -143,12 +130,19 @@ static void registerUnregisteredSymbol(Symbol &sym) {
 }
 
 static void writeRpn(std::vector<uint8_t> &rpnexpr, std::vector<uint8_t> const &rpn) {
-	std::string symName;
 	size_t rpnptr = 0;
 
 	for (size_t offset = 0; offset < rpn.size();) {
 		uint8_t rpndata = rpn[offset++];
 
+		auto getSymName = [&]() {
+			std::string symName;
+			for (uint8_t c; (c = rpn[offset++]) != 0;) {
+				symName += c;
+			}
+			return symName;
+		};
+
 		switch (rpndata) {
 			Symbol *sym;
 			uint32_t value;
@@ -163,17 +157,8 @@ static void writeRpn(std::vector<uint8_t> &rpnexpr, std::vector<uint8_t> const &
 			break;
 
 		case RPN_SYM:
-			symName.clear();
-			for (;;) {
-				uint8_t c = rpn[offset++];
-				if (c == 0) {
-					break;
-				}
-				symName += c;
-			}
-
 			// The symbol name is always written expanded
-			sym = sym_FindExactSymbol(symName);
+			sym = sym_FindExactSymbol(getSymName());
 			if (sym->isConstant()) {
 				rpnexpr[rpnptr++] = RPN_CONST;
 				value = sym->getConstantValue();
@@ -190,17 +175,8 @@ static void writeRpn(std::vector<uint8_t> &rpnexpr, std::vector<uint8_t> const &
 			break;
 
 		case RPN_BANK_SYM:
-			symName.clear();
-			for (;;) {
-				uint8_t c = rpn[offset++];
-				if (c == 0) {
-					break;
-				}
-				symName += c;
-			}
-
 			// The symbol name is always written expanded
-			sym = sym_FindExactSymbol(symName);
+			sym = sym_FindExactSymbol(getSymName());
 			registerUnregisteredSymbol(*sym); // Ensure that `sym->ID` is set
 			value = sym->ID;
 
@@ -269,7 +245,8 @@ static void initPatch(Patch &patch, uint32_t type, Expression const &expr, uint3
 
 void out_CreatePatch(uint32_t type, Expression const &expr, uint32_t ofs, uint32_t pcShift) {
 	// Add the patch to the list
-	Patch &patch = currentSection->patches.emplace_front();
+	assume(sect_GetOutputBank().has_value());
+	Patch &patch = *sect_AddOutputPatch();
 
 	initPatch(patch, type, expr, ofs);
 
@@ -311,20 +288,24 @@ static void writeFileStackNode(FileStackNode const &node, FILE *file) {
 }
 
 void out_WriteObject() {
-	if (objectFileName.empty()) {
+	if (options.objectFileName.empty()) {
 		return;
 	}
 
-	FILE *file;
-	if (objectFileName != "-") {
-		file = fopen(objectFileName.c_str(), "wb");
+	static FILE *file; // `static` so `sect_ForEach` callback can see it
+	if (options.objectFileName != "-") {
+		file = fopen(options.objectFileName.c_str(), "wb");
 	} else {
-		objectFileName = "<stdout>";
+		options.objectFileName = "<stdout>";
 		(void)setmode(STDOUT_FILENO, O_BINARY);
 		file = stdout;
 	}
 	if (!file) {
-		err("Failed to open object file '%s'", objectFileName.c_str()); // LCOV_EXCL_LINE
+		// LCOV_EXCL_START
+		fatal(
+		    "Failed to open object file '%s': %s", options.objectFileName.c_str(), strerror(errno)
+		);
+		// LCOV_EXCL_STOP
 	}
 	Defer closeFile{[&] { fclose(file); }};
 
@@ -335,10 +316,10 @@ void out_WriteObject() {
 	putLong(RGBDS_OBJECT_REV, file);
 
 	putLong(objectSymbols.size(), file);
-	putLong(sectionList.size(), file);
+	putLong(sect_CountSections(), file);
 
 	putLong(fileStackNodes.size(), file);
-	for (auto it = fileStackNodes.begin(); it != fileStackNodes.end(); it++) {
+	for (auto it = fileStackNodes.begin(); it != fileStackNodes.end(); ++it) {
 		FileStackNode const &node = **it;
 
 		writeFileStackNode(node, file);
@@ -351,9 +332,7 @@ void out_WriteObject() {
 		writeSymbol(*sym, file);
 	}
 
-	for (Section const &sect : sectionList) {
-		writeSection(sect, file);
-	}
+	sect_ForEach([](Section &sect) { writeSection(sect, file); });
 
 	putLong(assertions.size(), file);
 
@@ -362,18 +341,6 @@ void out_WriteObject() {
 	}
 }
 
-void out_SetFileName(std::string const &name) {
-	if (!objectFileName.empty()) {
-		warnx("Overriding output filename %s", objectFileName.c_str());
-	}
-	objectFileName = name;
-	// LCOV_EXCL_START
-	if (verbose) {
-		printf("Output filename %s\n", objectFileName.c_str());
-	}
-	// LCOV_EXCL_STOP
-}
-
 static void dumpString(std::string const &escape, FILE *file) {
 	for (char c : escape) {
 		// Escape characters that need escaping
@@ -402,6 +369,11 @@ static void dumpString(std::string const &escape, FILE *file) {
 	}
 }
 
+// Symbols are ordered by file, then by definition order
+static bool compareSymbols(Symbol const *sym1, Symbol const *sym2) {
+	return sym1->defIndex < sym2->defIndex;
+}
+
 static bool dumpEquConstants(FILE *file) {
 	static std::vector<Symbol *> equConstants; // `static` so `sym_ForEach` callback can see it
 	equConstants.clear();
@@ -411,10 +383,7 @@ static bool dumpEquConstants(FILE *file) {
 			equConstants.push_back(&sym);
 		}
 	});
-	// Constants are ordered by file, then by definition order
-	std::sort(RANGE(equConstants), [](Symbol *sym1, Symbol *sym2) -> bool {
-		return sym1->defIndex < sym2->defIndex;
-	});
+	std::sort(RANGE(equConstants), compareSymbols);
 
 	for (Symbol const *sym : equConstants) {
 		uint32_t value = static_cast<uint32_t>(sym->getOutputValue());
@@ -433,10 +402,7 @@ static bool dumpVariables(FILE *file) {
 			variables.push_back(&sym);
 		}
 	});
-	// Variables are ordered by file, then by definition order
-	std::sort(RANGE(variables), [](Symbol *sym1, Symbol *sym2) -> bool {
-		return sym1->defIndex < sym2->defIndex;
-	});
+	std::sort(RANGE(variables), compareSymbols);
 
 	for (Symbol const *sym : variables) {
 		uint32_t value = static_cast<uint32_t>(sym->getOutputValue());
@@ -455,10 +421,7 @@ static bool dumpEqusConstants(FILE *file) {
 			equsConstants.push_back(&sym);
 		}
 	});
-	// Constants are ordered by file, then by definition order
-	std::sort(RANGE(equsConstants), [](Symbol *sym1, Symbol *sym2) -> bool {
-		return sym1->defIndex < sym2->defIndex;
-	});
+	std::sort(RANGE(equsConstants), compareSymbols);
 
 	for (Symbol const *sym : equsConstants) {
 		fprintf(file, "def %s equs \"", sym->name.c_str());
@@ -497,13 +460,10 @@ static bool dumpMacros(FILE *file) {
 			macros.push_back(&sym);
 		}
 	});
-	// Macros are ordered by file, then by definition order
-	std::sort(RANGE(macros), [](Symbol *sym1, Symbol *sym2) -> bool {
-		return sym1->defIndex < sym2->defIndex;
-	});
+	std::sort(RANGE(macros), compareSymbols);
 
 	for (Symbol const *sym : macros) {
-		auto const &body = sym->getMacro();
+		ContentSpan const &body = sym->getMacro();
 		fprintf(file, "macro %s\n", sym->name.c_str());
 		fwrite(body.ptr.get(), 1, body.size, file);
 		fputs("endm\n", file);
@@ -524,7 +484,9 @@ void out_WriteState(std::string name, std::vector<StateFeature> const &features)
 		file = stdout;
 	}
 	if (!file) {
-		err("Failed to open state file '%s'", name.c_str()); // LCOV_EXCL_LINE
+		// LCOV_EXCL_START
+		fatal("Failed to open state file '%s': %s", name.c_str(), strerror(errno));
+		// LCOV_EXCL_STOP
 	}
 	Defer closeFile{[&] { fclose(file); }};
 
@@ -547,7 +509,7 @@ void out_WriteState(std::string name, std::vector<StateFeature> const &features)
 	for (StateFeature feature : features) {
 		fprintf(file, "\n; %s\n", dumpHeadings[feature]);
 		if (!dumpFuncs[feature](file)) {
-			fprintf(file, "; No values\n");
+			fputs("; No values\n", file);
 		}
 	}
 }

src/asm/rpn.cpp

@@ -4,6 +4,7 @@
 
 #include <inttypes.h>
 #include <limits.h>
+#include <optional>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -39,7 +40,7 @@ uint8_t *Expression::reserveSpace(uint32_t size, uint32_t patchSize) {
 
 int32_t Expression::getConstVal() const {
 	if (!isKnown()) {
-		error("Expected constant expression: %s\n", data.get<std::string>().c_str());
+		error("Expected constant expression: %s", std::get<std::string>(data).c_str());
 		return 0;
 	}
 	return value();
@@ -73,10 +74,10 @@ void Expression::makeNumber(uint32_t value) {
 void Expression::makeSymbol(std::string const &symName) {
 	clear();
 	if (Symbol *sym = sym_FindScopedSymbol(symName); sym_IsPC(sym) && !sect_GetSymbolSection()) {
-		error("PC has no value outside of a section\n");
+		error("PC has no value outside of a section");
 		data = 0;
 	} else if (sym && !sym->isNumeric() && !sym->isLabel()) {
-		error("'%s' is not a numeric symbol\n", symName.c_str());
+		error("'%s' is not a numeric symbol", symName.c_str());
 		data = 0;
 	} else if (!sym || !sym->isConstant()) {
 		isSymbol = true;
@@ -102,19 +103,19 @@ void Expression::makeBankSymbol(std::string const &symName) {
 	clear();
 	if (Symbol const *sym = sym_FindScopedSymbol(symName); sym_IsPC(sym)) {
 		// The @ symbol is treated differently.
-		if (!currentSection) {
-			error("PC has no bank outside of a section\n");
+		if (std::optional<uint32_t> outputBank = sect_GetOutputBank(); !outputBank) {
+			error("PC has no bank outside of a section");
 			data = 1;
-		} else if (currentSection->bank == UINT32_MAX) {
+		} else if (*outputBank == UINT32_MAX) {
 			data = "Current section's bank is not known";
 
 			*reserveSpace(1) = RPN_BANK_SELF;
 		} else {
-			data = static_cast<int32_t>(currentSection->bank);
+			data = static_cast<int32_t>(*outputBank);
 		}
 		return;
 	} else if (sym && !sym->isLabel()) {
-		error("BANK argument must be a label\n");
+		error("BANK argument must be a label");
 		data = 1;
 	} else {
 		sym = sym_Ref(symName);
@@ -325,9 +326,9 @@ void Expression::makeUnaryOp(RPNCommand op, Expression &&src) {
 		case RPN_TZCOUNT:
 			data = val != 0 ? ctz(uval) : 32;
 			break;
+		// LCOV_EXCL_START
 		default:
 			// `makeUnaryOp` should never be called with a non-unary operator!
-			// LCOV_EXCL_START
 			unreachable_();
 		}
 		// LCOV_EXCL_STOP
@@ -394,43 +395,37 @@ void Expression::makeBinaryOp(RPNCommand op, Expression &&src1, Expression const
 			break;
 		case RPN_SHL:
 			if (rval < 0) {
-				warning(
-				    WARNING_SHIFT_AMOUNT, "Shifting left by negative amount %" PRId32 "\n", rval
-				);
+				warning(WARNING_SHIFT_AMOUNT, "Shifting left by negative amount %" PRId32, rval);
 			}
 
 			if (rval >= 32) {
-				warning(WARNING_SHIFT_AMOUNT, "Shifting left by large amount %" PRId32 "\n", rval);
+				warning(WARNING_SHIFT_AMOUNT, "Shifting left by large amount %" PRId32, rval);
 			}
 
 			data = op_shift_left(lval, rval);
 			break;
 		case RPN_SHR:
 			if (lval < 0) {
-				warning(WARNING_SHIFT, "Shifting right negative value %" PRId32 "\n", lval);
+				warning(WARNING_SHIFT, "Shifting right negative value %" PRId32, lval);
 			}
 
 			if (rval < 0) {
-				warning(
-				    WARNING_SHIFT_AMOUNT, "Shifting right by negative amount %" PRId32 "\n", rval
-				);
+				warning(WARNING_SHIFT_AMOUNT, "Shifting right by negative amount %" PRId32, rval);
 			}
 
 			if (rval >= 32) {
-				warning(WARNING_SHIFT_AMOUNT, "Shifting right by large amount %" PRId32 "\n", rval);
+				warning(WARNING_SHIFT_AMOUNT, "Shifting right by large amount %" PRId32, rval);
 			}
 
 			data = op_shift_right(lval, rval);
 			break;
 		case RPN_USHR:
 			if (rval < 0) {
-				warning(
-				    WARNING_SHIFT_AMOUNT, "Shifting right by negative amount %" PRId32 "\n", rval
-				);
+				warning(WARNING_SHIFT_AMOUNT, "Shifting right by negative amount %" PRId32, rval);
 			}
 
 			if (rval >= 32) {
-				warning(WARNING_SHIFT_AMOUNT, "Shifting right by large amount %" PRId32 "\n", rval);
+				warning(WARNING_SHIFT_AMOUNT, "Shifting right by large amount %" PRId32, rval);
 			}
 
 			data = op_shift_right_unsigned(lval, rval);
@@ -440,13 +435,13 @@ void Expression::makeBinaryOp(RPNCommand op, Expression &&src1, Expression const
 			break;
 		case RPN_DIV:
 			if (rval == 0) {
-				fatalerror("Division by zero\n");
+				fatal("Division by zero");
 			}
 
 			if (lval == INT32_MIN && rval == -1) {
 				warning(
 				    WARNING_DIV,
-				    "Division of %" PRId32 " by -1 yields %" PRId32 "\n",
+				    "Division of %" PRId32 " by -1 yields %" PRId32,
 				    INT32_MIN,
 				    INT32_MIN
 				);
@@ -457,7 +452,7 @@ void Expression::makeBinaryOp(RPNCommand op, Expression &&src1, Expression const
 			break;
 		case RPN_MOD:
 			if (rval == 0) {
-				fatalerror("Modulo by zero\n");
+				fatal("Modulo by zero");
 			}
 
 			if (lval == INT32_MIN && rval == -1) {
@@ -468,14 +463,14 @@ void Expression::makeBinaryOp(RPNCommand op, Expression &&src1, Expression const
 			break;
 		case RPN_EXP:
 			if (rval < 0) {
-				fatalerror("Exponentiation by negative power\n");
+				fatal("Exponentiation by negative power");
 			}
 
 			data = op_exponent(lval, rval);
 			break;
+		// LCOV_EXCL_START
 		default:
 			// `makeBinaryOp` should never be called with a non-binary operator!
-			// LCOV_EXCL_START
 			unreachable_();
 		}
 		// LCOV_EXCL_STOP
@@ -551,7 +546,7 @@ bool Expression::makeCheckHRAM() {
 		// That range is valid, but deprecated
 		return true;
 	} else {
-		error("Source address $%" PRIx32 " not between $FF00 to $FFFF\n", val);
+		error("Source address $%" PRIx32 " not between $FF00 to $FFFF", val);
 	}
 	return false;
 }
@@ -561,7 +556,7 @@ void Expression::makeCheckRST() {
 		*reserveSpace(1) = RPN_RST;
 	} else if (int32_t val = value(); val & ~0x38) {
 		// A valid RST address must be masked with 0x38
-		error("Invalid address $%" PRIx32 " for RST\n", val);
+		error("Invalid address $%" PRIx32 " for RST", val);
 	}
 }
 
@@ -575,7 +570,7 @@ void Expression::makeCheckBitIndex(uint8_t mask) {
 	} else if (int32_t val = value(); val & ~0x07) {
 		// A valid bit index must be masked with 0x07
 		static char const *instructions[4] = {"instruction", "BIT", "RES", "SET"};
-		error("Invalid bit index %" PRId32 " for %s\n", val, instructions[mask >> 6]);
+		error("Invalid bit index %" PRId32 " for %s", val, instructions[mask >> 6]);
 	}
 }
 
@@ -593,20 +588,20 @@ bool checkNBit(int32_t v, uint8_t n, char const *name) {
 	if (v < -(1 << n) || v >= 1 << n) {
 		warning(
 		    WARNING_TRUNCATION_1,
-		    n == 8 && !name ? "%s must be %u-bit; use LOW() to force 8-bit\n"
-		                    : "%s must be %u-bit\n",
+		    "%s must be %u-bit%s",
 		    name ? name : "Expression",
-		    n
+		    n,
+		    n == 8 && !name ? "; use LOW() to force 8-bit" : ""
 		);
 		return false;
 	}
 	if (v < -(1 << (n - 1))) {
 		warning(
 		    WARNING_TRUNCATION_2,
-		    n == 8 && !name ? "%s must be %u-bit; use LOW() to force 8-bit\n"
-		                    : "%s must be %u-bit\n",
+		    "%s must be %u-bit%s",
 		    name ? name : "Expression",
-		    n
+		    n,
+		    n == 8 && !name ? "; use LOW() to force 8-bit" : ""
 		);
 		return false;
 	}

src/asm/section.cpp

@@ -5,7 +5,6 @@
 #include <algorithm>
 #include <errno.h>
 #include <inttypes.h>
-#include <optional>
 #include <stack>
 #include <stdio.h>
 #include <stdlib.h>
@@ -22,7 +21,7 @@
 #include "asm/symbol.hpp"
 #include "asm/warning.hpp"
 
-uint8_t fillByte;
+using namespace std::literals;
 
 struct UnionStackEntry {
 	uint32_t start;
@@ -38,29 +37,30 @@ struct SectionStackEntry {
 	std::stack<UnionStackEntry> unionStack;
 };
 
-std::stack<UnionStackEntry> currentUnionStack;
-std::deque<SectionStackEntry> sectionStack;
-std::deque<Section> sectionList;
-std::unordered_map<std::string, size_t> sectionMap; // Indexes into `sectionList`
-uint32_t curOffset; // Offset into the current section (see sect_GetSymbolOffset)
-Section *currentSection = nullptr;
+static Section *currentSection = nullptr;
+static std::deque<Section> sectionList;
+static std::unordered_map<std::string, size_t> sectionMap; // Indexes into `sectionList`
+
+static uint32_t curOffset; // Offset into the current section (see `sect_GetSymbolOffset`)
+
+static std::deque<SectionStackEntry> sectionStack;
+
 static Section *currentLoadSection = nullptr;
 static std::pair<Symbol const *, Symbol const *> currentLoadLabelScopes = {nullptr, nullptr};
-int32_t loadOffset; // Offset into the LOAD section's parent (see sect_GetOutputOffset)
+static int32_t loadOffset; // Offset into the LOAD section's parent (see sect_GetOutputOffset)
+
+static std::stack<UnionStackEntry> currentUnionStack;
 
-// A quick check to see if we have an initialized section
 [[nodiscard]]
 static bool requireSection() {
 	if (currentSection) {
 		return true;
 	}
 
-	error("Cannot output data outside of a SECTION\n");
+	error("Cannot output data outside of a SECTION");
 	return false;
 }
 
-// A quick check to see if we have an initialized section that can contain
-// this much initialized data
 [[nodiscard]]
 static bool requireCodeSection() {
 	if (!requireSection()) {
@@ -72,18 +72,26 @@ static bool requireCodeSection() {
 	}
 
 	error(
-	    "Section '%s' cannot contain code or data (not ROM0 or ROMX)\n",
-	    currentSection->name.c_str()
+	    "Section '%s' cannot contain code or data (not ROM0 or ROMX)", currentSection->name.c_str()
 	);
 	return false;
 }
 
+size_t sect_CountSections() {
+	return sectionList.size();
+}
+
+void sect_ForEach(void (*callback)(Section &)) {
+	for (Section &sect : sectionList) {
+		callback(sect);
+	}
+}
+
 void sect_CheckSizes() {
 	for (Section const &sect : sectionList) {
 		if (uint32_t maxSize = sectionTypeInfo[sect.type].size; sect.size > maxSize) {
 			error(
-			    "Section '%s' grew too big (max size = 0x%" PRIX32 " bytes, reached 0x%" PRIX32
-			    ")\n",
+			    "Section '%s' grew too big (max size = 0x%" PRIX32 " bytes, reached 0x%" PRIX32 ")",
 			    sect.name.c_str(),
 			    maxSize,
 			    sect.size
@@ -101,7 +109,7 @@ Section *sect_FindSectionByName(std::string const &name) {
 #define sectError(...) \
 	do { \
 		error(__VA_ARGS__); \
-		nbSectErrors++; \
+		++nbSectErrors; \
 	} while (0)
 
 static unsigned int mergeSectUnion(
@@ -113,18 +121,18 @@ static unsigned int mergeSectUnion(
 	// Unionized sections only need "compatible" constraints, and they end up with the strictest
 	// combination of both.
 	if (sect_HasData(type)) {
-		sectError("Cannot declare ROM sections as UNION\n");
+		sectError("Cannot declare ROM sections as UNION");
 	}
 
 	if (org != UINT32_MAX) {
 		// If both are fixed, they must be the same
 		if (sect.org != UINT32_MAX && sect.org != org) {
 			sectError(
-			    "Section already declared as fixed at different address $%04" PRIx32 "\n", sect.org
+			    "Section already declared as fixed at different address $%04" PRIx32, sect.org
 			);
 		} else if (sect.align != 0 && (mask(sect.align) & (org - sect.alignOfs))) {
 			sectError(
-			    "Section already declared as aligned to %u bytes (offset %" PRIu16 ")\n",
+			    "Section already declared as aligned to %u bytes (offset %" PRIu16 ")",
 			    1U << sect.align,
 			    sect.alignOfs
 			);
@@ -138,15 +146,14 @@ static unsigned int mergeSectUnion(
 		if (sect.org != UINT32_MAX) {
 			if ((sect.org - alignOffset) & mask(alignment)) {
 				sectError(
-				    "Section already declared as fixed at incompatible address $%04" PRIx32 "\n",
+				    "Section already declared as fixed at incompatible address $%04" PRIx32,
 				    sect.org
 				);
 			}
 			// Check if alignment offsets are compatible
 		} else if ((alignOffset & mask(sect.align)) != (sect.alignOfs & mask(alignment))) {
 			sectError(
-			    "Section already declared with incompatible %u"
-			    "-byte alignment (offset %" PRIu16 ")\n",
+			    "Section already declared with incompatible %u-byte alignment (offset %" PRIu16 ")",
 			    1U << sect.align,
 			    sect.alignOfs
 			);
@@ -174,12 +181,11 @@ static unsigned int
 		// If both are fixed, they must be the same
 		if (sect.org != UINT32_MAX && sect.org != curOrg) {
 			sectError(
-			    "Section already declared as fixed at incompatible address $%04" PRIx32 "\n",
-			    sect.org
+			    "Section already declared as fixed at incompatible address $%04" PRIx32, sect.org
 			);
 		} else if (sect.align != 0 && (mask(sect.align) & (curOrg - sect.alignOfs))) {
 			sectError(
-			    "Section already declared as aligned to %u bytes (offset %" PRIu16 ")\n",
+			    "Section already declared as aligned to %u bytes (offset %" PRIu16 ")",
 			    1U << sect.align,
 			    sect.alignOfs
 			);
@@ -199,15 +205,14 @@ static unsigned int
 		if (sect.org != UINT32_MAX) {
 			if ((sect.org - curOfs) & mask(alignment)) {
 				sectError(
-				    "Section already declared as fixed at incompatible address $%04" PRIx32 "\n",
+				    "Section already declared as fixed at incompatible address $%04" PRIx32,
 				    sect.org
 				);
 			}
 			// Check if alignment offsets are compatible
 		} else if ((curOfs & mask(sect.align)) != (sect.alignOfs & mask(alignment))) {
 			sectError(
-			    "Section already declared with incompatible %u"
-			    "-byte alignment (offset %" PRIu16 ")\n",
+			    "Section already declared with incompatible %u-byte alignment (offset %" PRIu16 ")",
 			    1U << sect.align,
 			    sect.alignOfs
 			);
@@ -234,12 +239,12 @@ static void mergeSections(
 
 	if (type != sect.type) {
 		sectError(
-		    "Section already exists but with type %s\n", sectionTypeInfo[sect.type].name.c_str()
+		    "Section already exists but with type %s", sectionTypeInfo[sect.type].name.c_str()
 		);
 	}
 
 	if (sect.modifier != mod) {
-		sectError("Section already declared as SECTION %s\n", sectionModNames[sect.modifier]);
+		sectError("Section already declared as SECTION %s", sectionModNames[sect.modifier]);
 	} else {
 		switch (mod) {
 		case SECTION_UNION:
@@ -256,21 +261,22 @@ static void mergeSections(
 			}
 			// If both specify a bank, it must be the same one
 			else if (bank != UINT32_MAX && sect.bank != bank) {
-				sectError("Section already declared with different bank %" PRIu32 "\n", sect.bank);
+				sectError("Section already declared with different bank %" PRIu32, sect.bank);
 			}
 			break;
 
 		case SECTION_NORMAL:
-			sectError("Section already defined previously at ");
+			sectError([&]() {
+				fputs("Section already defined previously at ", stderr);
 				sect.src->dump(sect.fileLine);
-			putc('\n', stderr);
+			});
 			break;
 		}
 	}
 
 	if (nbSectErrors) {
-		fatalerror(
-		    "Cannot create section \"%s\" (%u error%s)\n",
+		fatal(
+		    "Cannot create section \"%s\" (%u error%s)",
 		    sect.name.c_str(),
 		    nbSectErrors,
 		    nbSectErrors == 1 ? "" : "s"
@@ -280,7 +286,6 @@ static void mergeSections(
 
 #undef sectError
 
-// Create a new section, not yet in the list.
 static Section *createSection(
     std::string const &name,
     SectionType type,
@@ -315,7 +320,31 @@ static Section *createSection(
 	return &sect;
 }
 
-// Find a section by name and type. If it doesn't exist, create it.
+static Section *createSectionFragmentLiteral(Section const &parent) {
+	// Add the new section to the list, but do not update the map
+	Section &sect = sectionList.emplace_back();
+	assume(sectionMap.find(parent.name) != sectionMap.end());
+
+	sect.name = parent.name;
+	sect.type = parent.type;
+	sect.modifier = SECTION_FRAGMENT;
+	sect.src = fstk_GetFileStack();
+	sect.fileLine = lexer_GetLineNo();
+	sect.size = 0;
+	sect.org = UINT32_MAX;
+	sect.bank = parent.bank == 0 ? UINT32_MAX : parent.bank;
+	sect.align = 0;
+	sect.alignOfs = 0;
+
+	out_RegisterNode(sect.src);
+
+	// Section fragment literals must be ROM sections.
+	assume(sect_HasData(sect.type));
+	sect.data.resize(sectionTypeInfo[sect.type].size);
+
+	return &sect;
+}
+
 static Section *getSection(
     std::string const &name,
     SectionType type,
@@ -332,11 +361,11 @@ static Section *getSection(
 	if (bank != UINT32_MAX) {
 		if (type != SECTTYPE_ROMX && type != SECTTYPE_VRAM && type != SECTTYPE_SRAM
 		    && type != SECTTYPE_WRAMX) {
-			error("BANK only allowed for ROMX, WRAMX, SRAM, or VRAM sections\n");
+			error("BANK only allowed for ROMX, WRAMX, SRAM, or VRAM sections");
 		} else if (bank < sectionTypeInfo[type].firstBank
 		           || bank > sectionTypeInfo[type].lastBank) {
 			error(
-			    "%s bank value $%04" PRIx32 " out of range ($%04" PRIx32 " to $%04" PRIx32 ")\n",
+			    "%s bank value $%04" PRIx32 " out of range ($%04" PRIx32 " to $%04" PRIx32 ")",
 			    sectionTypeInfo[type].name.c_str(),
 			    bank,
 			    sectionTypeInfo[type].firstBank,
@@ -350,7 +379,7 @@ static Section *getSection(
 
 	if (alignOffset >= 1 << alignment) {
 		error(
-		    "Alignment offset (%" PRIu16 ") must be smaller than alignment size (%u)\n",
+		    "Alignment offset (%" PRIu16 ") must be smaller than alignment size (%u)",
 		    alignOffset,
 		    1U << alignment
 		);
@@ -361,7 +390,7 @@ static Section *getSection(
 		if (org < sectionTypeInfo[type].startAddr || org > endaddr(type)) {
 			error(
 			    "Section \"%s\"'s fixed address $%04" PRIx32 " is outside of range [$%04" PRIx16
-			    "; $%04" PRIx16 "]\n",
+			    "; $%04" PRIx16 "]",
 			    name.c_str(),
 			    org,
 			    sectionTypeInfo[type].startAddr,
@@ -372,7 +401,7 @@ static Section *getSection(
 
 	if (alignment != 0) {
 		if (alignment > 16) {
-			error("Alignment must be between 0 and 16, not %u\n", alignment);
+			error("Alignment must be between 0 and 16, not %u", alignment);
 			alignment = 16;
 		}
 		// It doesn't make sense to have both alignment and org set
@@ -380,12 +409,12 @@ static Section *getSection(
 
 		if (org != UINT32_MAX) {
 			if ((org - alignOffset) & mask) {
-				error("Section \"%s\"'s fixed address doesn't match its alignment\n", name.c_str());
+				error("Section \"%s\"'s fixed address doesn't match its alignment", name.c_str());
 			}
 			alignment = 0; // Ignore it if it's satisfied
 		} else if (sectionTypeInfo[type].startAddr & mask) {
 			error(
-			    "Section \"%s\"'s alignment cannot be attained in %s\n",
+			    "Section \"%s\"'s alignment cannot be attained in %s",
 			    name.c_str(),
 			    sectionTypeInfo[type].name.c_str()
 			);
@@ -412,15 +441,24 @@ static Section *getSection(
 	return sect;
 }
 
-// Set the current section
 static void changeSection() {
 	if (!currentUnionStack.empty()) {
-		fatalerror("Cannot change the section within a UNION\n");
+		fatal("Cannot change the section within a UNION");
 	}
 
 	sym_ResetCurrentLabelScopes();
 }
 
+uint32_t Section::getID() const {
+	// Section fragments share the same name but have different IDs, so search by identity
+	if (auto search =
+	        std::find_if(RANGE(sectionList), [this](Section const &s) { return &s == this; });
+	    search != sectionList.end()) {
+		return static_cast<uint32_t>(std::distance(sectionList.begin(), search));
+	}
+	return UINT32_MAX; // LCOV_EXCL_LINE
+}
+
 bool Section::isSizeKnown() const {
 	// SECTION UNION and SECTION FRAGMENT can still grow
 	if (modifier != SECTION_NORMAL) {
@@ -442,7 +480,6 @@ bool Section::isSizeKnown() const {
 	return true;
 }
 
-// Set the current section by name and type
 void sect_NewSection(
     std::string const &name,
     SectionType type,
@@ -452,7 +489,7 @@ void sect_NewSection(
 ) {
 	for (SectionStackEntry &entry : sectionStack) {
 		if (entry.section && entry.section->name == name) {
-			fatalerror("Section '%s' is already on the stack\n", name.c_str());
+			fatal("Section '%s' is already on the stack", name.c_str());
 		}
 	}
 
@@ -468,7 +505,6 @@ void sect_NewSection(
 	currentSection = sect;
 }
 
-// Set the current section by name and type
 void sect_SetLoadSection(
     std::string const &name,
     SectionType type,
@@ -486,7 +522,7 @@ void sect_SetLoadSection(
 	}
 
 	if (sect_HasData(type)) {
-		error("`LOAD` blocks cannot create a ROM section\n");
+		error("`LOAD` blocks cannot create a ROM section");
 		return;
 	}
 
@@ -505,13 +541,11 @@ void sect_SetLoadSection(
 
 void sect_EndLoadSection(char const *cause) {
 	if (cause) {
-		warning(
-		    WARNING_UNTERMINATED_LOAD, "`LOAD` block without `ENDL` terminated by `%s`\n", cause
-		);
+		warning(WARNING_UNTERMINATED_LOAD, "`LOAD` block without `ENDL` terminated by `%s`", cause);
 	}
 
 	if (!currentLoadSection) {
-		error("Found `ENDL` outside of a `LOAD` block\n");
+		error("Found `ENDL` outside of a `LOAD` block");
 		return;
 	}
 
@@ -524,7 +558,7 @@ void sect_EndLoadSection(char const *cause) {
 
 void sect_CheckLoadClosed() {
 	if (currentLoadSection) {
-		warning(WARNING_UNTERMINATED_LOAD, "`LOAD` block without `ENDL` terminated by EOF\n");
+		warning(WARNING_UNTERMINATED_LOAD, "`LOAD` block without `ENDL` terminated by EOF");
 	}
 }
 
@@ -532,7 +566,6 @@ Section *sect_GetSymbolSection() {
 	return currentLoadSection ? currentLoadSection : currentSection;
 }
 
-// The offset into the section above
 uint32_t sect_GetSymbolOffset() {
 	return curOffset;
 }
@@ -541,6 +574,14 @@ uint32_t sect_GetOutputOffset() {
 	return curOffset + loadOffset;
 }
 
+std::optional<uint32_t> sect_GetOutputBank() {
+	return currentSection ? std::optional<uint32_t>(currentSection->bank) : std::nullopt;
+}
+
+Patch *sect_AddOutputPatch() {
+	return currentSection ? &currentSection->patches.emplace_front() : nullptr;
+}
+
 // Returns how many bytes need outputting for the specified alignment and offset to succeed
 uint32_t sect_GetAlignBytes(uint8_t alignment, uint16_t offset) {
 	Section *sect = sect_GetSymbolSection();
@@ -572,24 +613,36 @@ void sect_AlignPC(uint8_t alignment, uint16_t offset) {
 	uint32_t alignSize = 1 << alignment; // Size of an aligned "block"
 
 	if (sect->org != UINT32_MAX) {
-		if ((sect->org + curOffset - offset) % alignSize) {
+		if (uint32_t actualOffset = (sect->org + curOffset) % alignSize; actualOffset != offset) {
 			error(
-			    "Section's fixed address fails required alignment (PC = $%04" PRIx32 ")\n",
-			    sect->org + curOffset
+			    "Section is misaligned (at PC = $%04" PRIx32 ", expected ALIGN[%" PRIu32
+			    ", %" PRIu32 "], got ALIGN[%" PRIu32 ", %" PRIu32 "])",
+			    sect->org + curOffset,
+			    alignment,
+			    offset,
+			    alignment,
+			    actualOffset
 			);
 		}
-	} else if (sect->align != 0
-	           && (((sect->alignOfs + curOffset) % (1u << sect->align)) - offset) % alignSize) {
+	} else {
+		if (uint32_t actualOffset = (sect->alignOfs + curOffset) % alignSize,
+		    sectAlignSize = 1 << sect->align;
+		    sect->align != 0 && actualOffset % sectAlignSize != offset % sectAlignSize) {
 			error(
-		    "Section's alignment fails required alignment (offset from section start = $%04" PRIx32
-		    ")\n",
-		    curOffset
+			    "Section is misaligned ($%04" PRIx32
+			    " bytes into the section, expected ALIGN[%" PRIu32 ", %" PRIu32
+			    "], got ALIGN[%" PRIu32 ", %" PRIu32 "])",
+			    curOffset,
+			    alignment,
+			    offset,
+			    alignment,
+			    actualOffset
 			);
 		} else if (alignment >= 16) {
 			// Treat an alignment large enough as fixing the address.
 			// Note that this also ensures that a section's alignment never becomes 16 or greater.
 			if (alignment > 16) {
-			error("Alignment must be between 0 and 16, not %u\n", alignment);
+				error("Alignment must be between 0 and 16, not %u", alignment);
 			}
 			sect->align = 0; // Reset the alignment, since we're fixing the address.
 			sect->org = offset - curOffset;
@@ -598,11 +651,12 @@ void sect_AlignPC(uint8_t alignment, uint16_t offset) {
 			// We need `(sect->alignOfs + curOffset) % alignSize == offset`
 			sect->alignOfs = (offset - curOffset) % alignSize;
 		}
+	}
 }
 
 static void growSection(uint32_t growth) {
 	if (growth > 0 && curOffset > UINT32_MAX - growth) {
-		fatalerror("Section size would overflow internal counter\n");
+		fatal("Section size would overflow internal counter");
 	}
 	curOffset += growth;
 	if (uint32_t outOffset = sect_GetOutputOffset(); outOffset > currentSection->size) {
@@ -643,11 +697,11 @@ void sect_StartUnion() {
 	// your own peril! ^^
 
 	if (!currentSection) {
-		error("UNIONs must be inside a SECTION\n");
+		error("UNIONs must be inside a SECTION");
 		return;
 	}
 	if (sect_HasData(currentSection->type)) {
-		error("Cannot use UNION inside of ROM0 or ROMX sections\n");
+		error("Cannot use UNION inside of ROM0 or ROMX sections");
 		return;
 	}
 
@@ -666,7 +720,7 @@ static void endUnionMember() {
 
 void sect_NextUnionMember() {
 	if (currentUnionStack.empty()) {
-		error("Found NEXTU outside of a UNION construct\n");
+		error("Found NEXTU outside of a UNION construct");
 		return;
 	}
 	endUnionMember();
@@ -674,7 +728,7 @@ void sect_NextUnionMember() {
 
 void sect_EndUnion() {
 	if (currentUnionStack.empty()) {
-		error("Found ENDU outside of a UNION construct\n");
+		error("Found ENDU outside of a UNION construct");
 		return;
 	}
 	endUnionMember();
@@ -684,11 +738,10 @@ void sect_EndUnion() {
 
 void sect_CheckUnionClosed() {
 	if (!currentUnionStack.empty()) {
-		error("Unterminated UNION construct\n");
+		error("Unterminated UNION construct");
 	}
 }
 
-// Output a constant byte
 void sect_ConstByte(uint8_t byte) {
 	if (!requireCodeSection()) {
 		return;
@@ -697,52 +750,48 @@ void sect_ConstByte(uint8_t byte) {
 	writeByte(byte);
 }
 
-// Output a string's character units as bytes
-void sect_ByteString(std::vector<int32_t> const &string) {
+void sect_ByteString(std::vector<int32_t> const &str) {
 	if (!requireCodeSection()) {
 		return;
 	}
 
-	for (int32_t unit : string) {
+	for (int32_t unit : str) {
 		if (!checkNBit(unit, 8, "All character units")) {
 			break;
 		}
 	}
 
-	for (int32_t unit : string) {
+	for (int32_t unit : str) {
 		writeByte(static_cast<uint8_t>(unit));
 	}
 }
 
-// Output a string's character units as words
-void sect_WordString(std::vector<int32_t> const &string) {
+void sect_WordString(std::vector<int32_t> const &str) {
 	if (!requireCodeSection()) {
 		return;
 	}
 
-	for (int32_t unit : string) {
+	for (int32_t unit : str) {
 		if (!checkNBit(unit, 16, "All character units")) {
 			break;
 		}
 	}
 
-	for (int32_t unit : string) {
+	for (int32_t unit : str) {
 		writeWord(static_cast<uint16_t>(unit));
 	}
 }
 
-// Output a string's character units as longs
-void sect_LongString(std::vector<int32_t> const &string) {
+void sect_LongString(std::vector<int32_t> const &str) {
 	if (!requireCodeSection()) {
 		return;
 	}
 
-	for (int32_t unit : string) {
+	for (int32_t unit : str) {
 		writeLong(static_cast<uint32_t>(unit));
 	}
 }
 
-// Skip this many bytes
 void sect_Skip(uint32_t skip, bool ds) {
 	if (!requireSection()) {
 		return;
@@ -754,7 +803,7 @@ void sect_Skip(uint32_t skip, bool ds) {
 		if (!ds) {
 			warning(
 			    WARNING_EMPTY_DATA_DIRECTIVE,
-			    "%s directive without data in ROM\n",
+			    "%s directive without data in ROM",
 			    (skip == 4)   ? "DL"
 			    : (skip == 2) ? "DW"
 			                  : "DB"
@@ -762,12 +811,11 @@ void sect_Skip(uint32_t skip, bool ds) {
 		}
 		// We know we're in a code SECTION
 		while (skip--) {
-			writeByte(fillByte);
+			writeByte(options.padByte);
 		}
 	}
 }
 
-// Output a byte that can be relocatable or constant
 void sect_RelByte(Expression const &expr, uint32_t pcShift) {
 	if (!requireCodeSection()) {
 		return;
@@ -781,13 +829,12 @@ void sect_RelByte(Expression const &expr, uint32_t pcShift) {
 	}
 }
 
-// Output several bytes that can be relocatable or constant
 void sect_RelBytes(uint32_t n, std::vector<Expression> const &exprs) {
 	if (!requireCodeSection()) {
 		return;
 	}
 
-	for (uint32_t i = 0; i < n; i++) {
+	for (uint32_t i = 0; i < n; ++i) {
 		if (Expression const &expr = exprs[i % exprs.size()]; !expr.isKnown()) {
 			createPatch(PATCHTYPE_BYTE, expr, i);
 			writeByte(0);
@@ -797,7 +844,6 @@ void sect_RelBytes(uint32_t n, std::vector<Expression> const &exprs) {
 	}
 }
 
-// Output a word that can be relocatable or constant
 void sect_RelWord(Expression const &expr, uint32_t pcShift) {
 	if (!requireCodeSection()) {
 		return;
@@ -811,7 +857,6 @@ void sect_RelWord(Expression const &expr, uint32_t pcShift) {
 	}
 }
 
-// Output a long that can be relocatable or constant
 void sect_RelLong(Expression const &expr, uint32_t pcShift) {
 	if (!requireCodeSection()) {
 		return;
@@ -825,7 +870,6 @@ void sect_RelLong(Expression const &expr, uint32_t pcShift) {
 	}
 }
 
-// Output a PC-relative byte that can be relocatable or constant
 void sect_PCRelByte(Expression const &expr, uint32_t pcShift) {
 	if (!requireCodeSection()) {
 		return;
@@ -849,7 +893,7 @@ void sect_PCRelByte(Expression const &expr, uint32_t pcShift) {
 		if (offset < -128 || offset > 127) {
 			error(
 			    "JR target must be between -128 and 127 bytes away, not %" PRId16
-			    "; use JP instead\n",
+			    "; use JP instead",
 			    offset
 			);
 			writeByte(0);
@@ -859,14 +903,9 @@ void sect_PCRelByte(Expression const &expr, uint32_t pcShift) {
 	}
 }
 
-// Output a binary file
-void sect_BinaryFile(std::string const &name, int32_t startPos) {
-	if (startPos < 0) {
-		error("Start position cannot be negative (%" PRId32 ")\n", startPos);
-		startPos = 0;
-	}
+bool sect_BinaryFile(std::string const &name, uint32_t startPos) {
 	if (!requireCodeSection()) {
-		return;
+		return false;
 	}
 
 	FILE *file = nullptr;
@@ -874,40 +913,26 @@ void sect_BinaryFile(std::string const &name, int32_t startPos) {
 		file = fopen(fullPath->c_str(), "rb");
 	}
 	if (!file) {
-		if (generatedMissingIncludes) {
-			// LCOV_EXCL_START
-			if (verbose) {
-				printf("Aborting (-MG) on INCBIN file '%s' (%s)\n", name.c_str(), strerror(errno));
-			}
-			// LCOV_EXCL_STOP
-			failedOnMissingInclude = true;
-		} else {
-			error("Error opening INCBIN file '%s': %s\n", name.c_str(), strerror(errno));
-		}
-		return;
+		return fstk_FileError(name, "INCBIN");
 	}
 	Defer closeFile{[&] { fclose(file); }};
 
-	if (fseek(file, 0, SEEK_END) != -1) {
+	if (fseek(file, 0, SEEK_END) == 0) {
 		if (startPos > ftell(file)) {
-			error("Specified start position is greater than length of file '%s'\n", name.c_str());
-			return;
+			error("Specified start position is greater than length of file '%s'", name.c_str());
+			return false;
 		}
 		// The file is seekable; skip to the specified start position
 		fseek(file, startPos, SEEK_SET);
 	} else {
 		if (errno != ESPIPE) {
-			error(
-			    "Error determining size of INCBIN file '%s': %s\n", name.c_str(), strerror(errno)
-			);
+			error("Error determining size of INCBIN file '%s': %s", name.c_str(), strerror(errno));
 		}
 		// The file isn't seekable, so we'll just skip bytes one at a time
 		while (startPos--) {
 			if (fgetc(file) == EOF) {
-				error(
-				    "Specified start position is greater than length of file '%s'\n", name.c_str()
-				);
-				return;
+				error("Specified start position is greater than length of file '%s'", name.c_str());
+				return false;
 			}
 		}
 	}
@@ -917,25 +942,17 @@ void sect_BinaryFile(std::string const &name, int32_t startPos) {
 	}
 
 	if (ferror(file)) {
-		error("Error reading INCBIN file '%s': %s\n", name.c_str(), strerror(errno));
+		error("Error reading INCBIN file '%s': %s", name.c_str(), strerror(errno));
 	}
+	return false;
 }
 
-// Output a slice of a binary file
-void sect_BinaryFileSlice(std::string const &name, int32_t startPos, int32_t length) {
-	if (startPos < 0) {
-		error("Start position cannot be negative (%" PRId32 ")\n", startPos);
-		startPos = 0;
-	}
-	if (length < 0) {
-		error("Number of bytes to read cannot be negative (%" PRId32 ")\n", length);
-		length = 0;
-	}
+bool sect_BinaryFileSlice(std::string const &name, uint32_t startPos, uint32_t length) {
 	if (!requireCodeSection()) {
-		return;
+		return false;
 	}
 	if (length == 0) { // Don't even bother with 0-byte slices
-		return;
+		return false;
 	}
 
 	FILE *file = nullptr;
@@ -943,50 +960,36 @@ void sect_BinaryFileSlice(std::string const &name, int32_t startPos, int32_t len
 		file = fopen(fullPath->c_str(), "rb");
 	}
 	if (!file) {
-		if (generatedMissingIncludes) {
-			// LCOV_EXCL_START
-			if (verbose) {
-				printf("Aborting (-MG) on INCBIN file '%s' (%s)\n", name.c_str(), strerror(errno));
-			}
-			// LCOV_EXCL_STOP
-			failedOnMissingInclude = true;
-		} else {
-			error("Error opening INCBIN file '%s': %s\n", name.c_str(), strerror(errno));
-		}
-		return;
+		return fstk_FileError(name, "INCBIN");
 	}
 	Defer closeFile{[&] { fclose(file); }};
 
-	if (fseek(file, 0, SEEK_END) != -1) {
-		if (int32_t fsize = ftell(file); startPos > fsize) {
-			error("Specified start position is greater than length of file '%s'\n", name.c_str());
-			return;
+	if (fseek(file, 0, SEEK_END) == 0) {
+		if (long fsize = ftell(file); startPos > fsize) {
+			error("Specified start position is greater than length of file '%s'", name.c_str());
+			return false;
 		} else if (startPos + length > fsize) {
 			error(
 			    "Specified range in INCBIN file '%s' is out of bounds (%" PRIu32 " + %" PRIu32
-			    " > %" PRIu32 ")\n",
+			    " > %ld)",
 			    name.c_str(),
 			    startPos,
 			    length,
 			    fsize
 			);
-			return;
+			return false;
 		}
 		// The file is seekable; skip to the specified start position
 		fseek(file, startPos, SEEK_SET);
 	} else {
 		if (errno != ESPIPE) {
-			error(
-			    "Error determining size of INCBIN file '%s': %s\n", name.c_str(), strerror(errno)
-			);
+			error("Error determining size of INCBIN file '%s': %s", name.c_str(), strerror(errno));
 		}
 		// The file isn't seekable, so we'll just skip bytes one at a time
 		while (startPos--) {
 			if (fgetc(file) == EOF) {
-				error(
-				    "Specified start position is greater than length of file '%s'\n", name.c_str()
-				);
-				return;
+				error("Specified start position is greater than length of file '%s'", name.c_str());
+				return false;
 			}
 		}
 	}
@@ -995,18 +998,18 @@ void sect_BinaryFileSlice(std::string const &name, int32_t startPos, int32_t len
 		if (int byte = fgetc(file); byte != EOF) {
 			writeByte(byte);
 		} else if (ferror(file)) {
-			error("Error reading INCBIN file '%s': %s\n", name.c_str(), strerror(errno));
+			error("Error reading INCBIN file '%s': %s", name.c_str(), strerror(errno));
 		} else {
 			error(
-			    "Premature end of INCBIN file '%s' (%" PRId32 " bytes left to read)\n",
+			    "Premature end of INCBIN file '%s' (%" PRId32 " bytes left to read)",
 			    name.c_str(),
 			    length + 1
 			);
 		}
 	}
+	return false;
 }
 
-// Section stack routines
 void sect_PushSection() {
 	sectionStack.push_front({
 	    .section = currentSection,
@@ -1026,7 +1029,7 @@ void sect_PushSection() {
 
 void sect_PopSection() {
 	if (sectionStack.empty()) {
-		fatalerror("No entries in the section stack\n");
+		fatal("No entries in the section stack");
 	}
 
 	if (currentLoadSection) {
@@ -1047,17 +1050,17 @@ void sect_PopSection() {
 
 void sect_CheckStack() {
 	if (!sectionStack.empty()) {
-		warning(WARNING_UNMATCHED_DIRECTIVE, "`PUSHS` without corresponding `POPS`\n");
+		warning(WARNING_UNMATCHED_DIRECTIVE, "`PUSHS` without corresponding `POPS`");
 	}
 }
 
 void sect_EndSection() {
 	if (!currentSection) {
-		fatalerror("Cannot end the section outside of a SECTION\n");
+		fatal("Cannot end the section outside of a SECTION");
 	}
 
 	if (!currentUnionStack.empty()) {
-		fatalerror("Cannot end the section within a UNION\n");
+		fatal("Cannot end the section within a UNION");
 	}
 
 	if (currentLoadSection) {
@@ -1068,3 +1071,40 @@ void sect_EndSection() {
 	currentSection = nullptr;
 	sym_ResetCurrentLabelScopes();
 }
+
+std::string sect_PushSectionFragmentLiteral() {
+	static uint64_t nextFragmentLiteralID = 0;
+
+	// Like `requireCodeSection` but fatal
+	if (!currentSection) {
+		fatal("Cannot output fragment literals outside of a SECTION");
+	}
+	if (!sect_HasData(currentSection->type)) {
+		fatal(
+		    "Section '%s' cannot contain fragment literals (not ROM0 or ROMX)",
+		    currentSection->name.c_str()
+		);
+	}
+
+	if (currentLoadSection) {
+		fatal("`LOAD` blocks cannot contain fragment literals");
+	}
+	if (currentSection->modifier == SECTION_UNION) {
+		fatal("`SECTION UNION` cannot contain fragment literals");
+	}
+
+	// A section containing a fragment literal has to become a fragment too
+	currentSection->modifier = SECTION_FRAGMENT;
+
+	Section *parent = currentSection;
+	sect_PushSection(); // Resets `currentSection`
+
+	Section *sect = createSectionFragmentLiteral(*parent);
+
+	changeSection();
+	curOffset = sect->size;
+	currentSection = sect;
+
+	// Return a symbol ID to use for the address of this section fragment
+	return "$"s + std::to_string(nextFragmentLiteralID++);
+}

src/asm/symbol.cpp

@@ -3,12 +3,13 @@
 #include "asm/symbol.hpp"
 
 #include <algorithm>
+#include <errno.h>
 #include <inttypes.h>
 #include <stdio.h>
 #include <unordered_map>
 #include <unordered_set>
 
-#include "error.hpp"
+#include "diagnostics.hpp"
 #include "helpers.hpp" // assume
 #include "util.hpp"
 #include "version.hpp"
@@ -21,21 +22,24 @@
 
 using namespace std::literals;
 
-std::unordered_map<std::string, Symbol> symbols;
-std::unordered_set<std::string> purgedSymbols;
+static std::unordered_map<std::string, Symbol> symbols;
+static std::unordered_set<std::string> purgedSymbols;
 
 static Symbol const *globalScope = nullptr; // Current section's global label scope
 static Symbol const *localScope = nullptr;  // Current section's local label scope
+
 static Symbol *PCSymbol;
 static Symbol *NARGSymbol;
 static Symbol *globalScopeSymbol;
 static Symbol *localScopeSymbol;
 static Symbol *RSSymbol;
+
 static char savedTIME[256];
 static char savedDATE[256];
 static char savedTIMESTAMP_ISO8601_LOCAL[256];
 static char savedTIMESTAMP_ISO8601_UTC[256];
-static bool exportAll;
+
+static bool exportAll = false; // -E
 
 bool sym_IsPC(Symbol const *sym) {
 	return sym == PCSymbol;
@@ -51,14 +55,14 @@ static int32_t NARGCallback() {
 	if (MacroArgs const *macroArgs = fstk_GetCurrentMacroArgs(); macroArgs) {
 		return macroArgs->nbArgs();
 	} else {
-		error("_NARG has no value outside of a macro\n");
+		error("_NARG has no value outside of a macro");
 		return 0;
 	}
 }
 
 static std::shared_ptr<std::string> globalScopeCallback() {
 	if (!globalScope) {
-		error("\".\" has no value outside of a label scope\n");
+		error("\".\" has no value outside of a label scope");
 		return std::make_shared<std::string>("");
 	}
 	return std::make_shared<std::string>(globalScope->name);
@@ -66,7 +70,7 @@ static std::shared_ptr<std::string> globalScopeCallback() {
 
 static std::shared_ptr<std::string> localScopeCallback() {
 	if (!localScope) {
-		error("\"..\" has no value outside of a local label scope\n");
+		error("\"..\" has no value outside of a local label scope");
 		return std::make_shared<std::string>("");
 	}
 	return std::make_shared<std::string>(localScope->name);
@@ -111,13 +115,13 @@ std::shared_ptr<std::string> Symbol::getEqus() const {
 }
 
 static void dumpFilename(Symbol const &sym) {
+	fputs(" at ", stderr);
 	if (sym.src) {
 		sym.src->dump(sym.fileLine);
-		putc('\n', stderr);
 	} else if (sym.isBuiltin) {
-		fputs("<builtin>\n", stderr);
+		fputs("<builtin>", stderr);
 	} else {
-		fputs("<command-line>\n", stderr);
+		fputs("<command-line>", stderr);
 	}
 }
 
@@ -140,23 +144,25 @@ static bool isValidIdentifier(std::string const &s) {
 static void alreadyDefinedError(Symbol const &sym, char const *asType) {
 	if (sym.isBuiltin && !sym_FindScopedValidSymbol(sym.name)) {
 		// `DEF()` would return false, so we should not claim the symbol is already defined
-		error("'%s' is reserved for a built-in symbol\n", sym.name.c_str());
+		error("'%s' is reserved for a built-in symbol", sym.name.c_str());
 	} else {
-		error("'%s' already defined", sym.name.c_str());
+		error([&]() {
+			fprintf(stderr, "'%s' already defined", sym.name.c_str());
 			if (asType) {
 				fprintf(stderr, " as %s", asType);
 			}
-		fputs(" at ", stderr);
 			dumpFilename(sym);
-		if (sym.type == SYM_EQUS) {
+			if (sym.type != SYM_EQUS) {
+				return;
+			}
 			if (std::string const &contents = *sym.getEqus(); isValidIdentifier(contents)) {
 				fprintf(
 				    stderr,
-				    "    (should it be {interpolated} to define its contents \"%s\"?)\n",
+				    "\n    (should it be {interpolated} to define its contents \"%s\"?)",
 				    contents.c_str()
 				);
 			}
-		}
+		});
 	}
 }
 
@@ -164,9 +170,9 @@ static void redefinedError(Symbol const &sym) {
 	assume(sym.isBuiltin);
 	if (!sym_FindScopedValidSymbol(sym.name)) {
 		// `DEF()` would return false, so we should not imply the symbol is already defined
-		error("'%s' is reserved for a built-in symbol\n", sym.name.c_str());
+		error("'%s' is reserved for a built-in symbol", sym.name.c_str());
 	} else {
-		error("Built-in symbol '%s' cannot be redefined\n", sym.name.c_str());
+		error("Built-in symbol '%s' cannot be redefined", sym.name.c_str());
 	}
 }
 
@@ -215,12 +221,12 @@ static bool isAutoScoped(std::string const &symName) {
 
 	// Check for nothing after the dot
 	if (dotPos == symName.length() - 1) {
-		fatalerror("'%s' is a nonsensical reference to an empty local label\n", symName.c_str());
+		fatal("'%s' is a nonsensical reference to an empty local label", symName.c_str());
 	}
 
 	// Check for more than one dot
 	if (symName.find('.', dotPos + 1) != std::string::npos) {
-		fatalerror("'%s' is a nonsensical reference to a nested local label\n", symName.c_str());
+		fatal("'%s' is a nonsensical reference to a nested local label", symName.c_str());
 	}
 
 	// Check for already-qualified local label
@@ -230,7 +236,7 @@ static bool isAutoScoped(std::string const &symName) {
 
 	// Check for unqualifiable local label
 	if (!globalScope) {
-		fatalerror("Unqualified local label '%s' in main scope\n", symName.c_str());
+		fatal("Unqualified local label '%s' in main scope", symName.c_str());
 	}
 
 	return true;
@@ -279,19 +285,19 @@ void sym_Purge(std::string const &symName) {
 
 	if (!sym) {
 		if (sym_IsPurgedScoped(symName)) {
-			error("'%s' was already purged\n", symName.c_str());
+			error("'%s' was already purged", symName.c_str());
 		} else {
-			error("'%s' not defined\n", symName.c_str());
+			error("'%s' not defined", symName.c_str());
 		}
 	} else if (sym->isBuiltin) {
-		error("Built-in symbol '%s' cannot be purged\n", symName.c_str());
+		error("Built-in symbol '%s' cannot be purged", symName.c_str());
 	} else if (sym->ID != UINT32_MAX) {
-		error("Symbol \"%s\" is referenced and thus cannot be purged\n", symName.c_str());
+		error("Symbol \"%s\" is referenced and thus cannot be purged", symName.c_str());
 	} else {
 		if (sym->isExported) {
-			warning(WARNING_PURGE_1, "Purging an exported symbol \"%s\"\n", symName.c_str());
+			warning(WARNING_PURGE_1, "Purging an exported symbol \"%s\"", symName.c_str());
 		} else if (sym->isLabel()) {
-			warning(WARNING_PURGE_2, "Purging a label \"%s\"\n", symName.c_str());
+			warning(WARNING_PURGE_2, "Purging a label \"%s\"", symName.c_str());
 		}
 		// Do not keep a reference to the label after purging it
 		if (sym == globalScope) {
@@ -331,12 +337,12 @@ uint32_t Symbol::getConstantValue() const {
 
 	if (sym_IsPC(this)) {
 		if (!getSection()) {
-			error("PC has no value outside of a section\n");
+			error("PC has no value outside of a section");
 		} else {
-			error("PC does not have a constant value; the current section is not fixed\n");
+			error("PC does not have a constant value; the current section is not fixed");
 		}
 	} else {
-		error("\"%s\" does not have a constant value\n", name.c_str());
+		error("\"%s\" does not have a constant value", name.c_str());
 	}
 	return 0;
 }
@@ -371,8 +377,10 @@ static Symbol *createNonrelocSymbol(std::string const &symName, bool numeric) {
 		return nullptr; // Don't allow overriding the symbol, that'd be bad!
 	} else if (!numeric) {
 		// The symbol has already been referenced, but it's not allowed
-		error("'%s' already referenced at ", symName.c_str());
+		error([&]() {
+			fprintf(stderr, "'%s' already referenced", symName.c_str());
 			dumpFilename(*sym);
+		});
 		return nullptr; // Don't allow overriding the symbol, that'd be bad!
 	}
 
@@ -437,8 +445,10 @@ Symbol *sym_RedefString(std::string const &symName, std::shared_ptr<std::string>
 		if (sym->isDefined()) {
 			alreadyDefinedError(*sym, "non-EQUS");
 		} else {
-			error("'%s' already referenced at ", symName.c_str());
+			error([&]() {
+				fprintf(stderr, "'%s' already referenced", symName.c_str());
 				dumpFilename(*sym);
+			});
 		}
 		return nullptr;
 	} else if (sym->isBuiltin) {
@@ -493,7 +503,7 @@ static Symbol *addLabel(std::string const &symName) {
 	sym->section = sect_GetSymbolSection();
 
 	if (sym && !sym->section) {
-		error("Label \"%s\" created outside of a SECTION\n", symName.c_str());
+		error("Label \"%s\" created outside of a SECTION", symName.c_str());
 	}
 
 	return sym;
@@ -538,7 +548,7 @@ Symbol *sym_AddAnonLabel() {
 	}
 
 	std::string anon = sym_MakeAnonLabelName(0, true); // The direction is important!
-	anonLabelID++;
+	++anonLabelID;
 	return addLabel(anon);
 }
 
@@ -549,7 +559,7 @@ std::string sym_MakeAnonLabelName(uint32_t ofs, bool neg) {
 		if (ofs > anonLabelID) {
 			error(
 			    "Reference to anonymous label %" PRIu32 " before, when only %" PRIu32
-			    " ha%s been created so far\n",
+			    " ha%s been created so far",
 			    ofs,
 			    anonLabelID,
 			    anonLabelID == 1 ? "s" : "ve"
@@ -558,12 +568,12 @@ std::string sym_MakeAnonLabelName(uint32_t ofs, bool neg) {
 			id = anonLabelID - ofs;
 		}
 	} else {
-		ofs--; // We're referencing symbols that haven't been created yet...
-		if (ofs > UINT32_MAX - anonLabelID) {
+		// We're referencing symbols that haven't been created yet...
+		if (--ofs > UINT32_MAX - anonLabelID) {
 			// LCOV_EXCL_START
 			error(
 			    "Reference to anonymous label %" PRIu32 " after, when only %" PRIu32
-			    " may still be created\n",
+			    " may still be created",
 			    ofs + 1,
 			    UINT32_MAX - anonLabelID
 			);
@@ -580,7 +590,7 @@ void sym_Export(std::string const &symName) {
 	if (symName.starts_with('!')) {
 		// LCOV_EXCL_START
 		// The parser does not accept anonymous labels for an `EXPORT` directive
-		error("Anonymous labels cannot be exported\n");
+		error("Anonymous labels cannot be exported");
 		return;
 		// LCOV_EXCL_STOP
 	}
@@ -666,7 +676,7 @@ void sym_Init(time_t now) {
 
 	// LCOV_EXCL_START
 	if (now == static_cast<time_t>(-1)) {
-		warn("Failed to determine current time");
+		warnx("Failed to determine current time: %s", strerror(errno));
 		// Fall back by pretending we are at the Epoch
 		now = 0;
 	}

src/asm/warning.cpp

@@ -9,7 +9,7 @@
 #include <stdlib.h>
 #include <string.h>
 
-#include "error.hpp"
+#include "diagnostics.hpp"
 #include "helpers.hpp"
 #include "itertools.hpp"
 
@@ -17,31 +17,14 @@
 #include "asm/lexer.hpp"
 #include "asm/main.hpp"
 
-unsigned int nbErrors = 0;
-unsigned int maxErrors = 0;
-
-Diagnostics warningStates;
-bool warningsAreErrors;
-
-enum WarningLevel {
-	LEVEL_DEFAULT,    // Warnings that are enabled by default
-	LEVEL_ALL,        // Warnings that probably indicate an error
-	LEVEL_EXTRA,      // Warnings that are less likely to indicate an error
-	LEVEL_EVERYTHING, // Literally every warning
-};
-
-struct WarningFlag {
-	char const *name;
-	WarningLevel level;
-};
-
-static WarningFlag const metaWarnings[] = {
+// clang-format off: nested initializers
+Diagnostics<WarningLevel, WarningID> warnings = {
+    .metaWarnings = {
         {"all",                  LEVEL_ALL       },
         {"extra",                LEVEL_EXTRA     },
         {"everything",           LEVEL_EVERYTHING},
-};
-
-static WarningFlag const warningFlags[NB_WARNINGS] = {
+    },
+    .warningFlags = {
         {"assert",               LEVEL_DEFAULT   },
         {"backwards-for",        LEVEL_ALL       },
         {"builtin-args",         LEVEL_ALL       },
@@ -68,245 +51,48 @@ static WarningFlag const warningFlags[NB_WARNINGS] = {
         {"truncation",           LEVEL_EXTRA     },
         {"unmapped-char",        LEVEL_DEFAULT   },
         {"unmapped-char",        LEVEL_ALL       },
-};
-
-static const struct {
-	WarningID firstID;
-	WarningID lastID;
-	uint8_t defaultLevel;
-} paramWarnings[] = {
+    },
+    .paramWarnings = {
         {WARNING_NUMERIC_STRING_1, WARNING_NUMERIC_STRING_2, 1},
         {WARNING_PURGE_1,          WARNING_PURGE_2,          1},
         {WARNING_TRUNCATION_1,     WARNING_TRUNCATION_2,     2},
         {WARNING_UNMAPPED_CHAR_1,  WARNING_UNMAPPED_CHAR_2,  1},
+    },
+    .state = DiagnosticsState<WarningID>(),
+    .nbErrors = 0,
 };
+// clang-format on
 
-enum WarningBehavior { DISABLED, ENABLED, ERROR };
-
-static WarningBehavior getWarningBehavior(WarningID id) {
-	// Check if warnings are globally disabled
-	if (!warnings) {
-		return WarningBehavior::DISABLED;
-	}
-
-	// Get the state of this warning flag
-	WarningState const &flagState = warningStates.flagStates[id];
-	WarningState const &metaState = warningStates.metaStates[id];
-
-	// If subsequent checks determine that the warning flag is enabled, this checks whether it has
-	// -Werror without -Wno-error=<flag> or -Wno-error=<meta>, which makes it into an error
-	bool warningIsError = warningsAreErrors && flagState.error != WARNING_DISABLED
-	                      && metaState.error != WARNING_DISABLED;
-	WarningBehavior enabledBehavior =
-	    warningIsError ? WarningBehavior::ERROR : WarningBehavior::ENABLED;
-
-	// First, check the state of the specific warning flag
-	if (flagState.state == WARNING_DISABLED) { // -Wno-<flag>
-		return WarningBehavior::DISABLED;
-	}
-	if (flagState.error == WARNING_ENABLED) { // -Werror=<flag>
-		return WarningBehavior::ERROR;
-	}
-	if (flagState.state == WARNING_ENABLED) { // -W<flag>
-		return enabledBehavior;
-	}
-
-	// If no flag is specified, check the state of the "meta" flags that affect this warning flag
-	if (metaState.state == WARNING_DISABLED) { // -Wno-<meta>
-		return WarningBehavior::DISABLED;
-	}
-	if (metaState.error == WARNING_ENABLED) { // -Werror=<meta>
-		return WarningBehavior::ERROR;
-	}
-	if (metaState.state == WARNING_ENABLED) { // -W<meta>
-		return enabledBehavior;
-	}
-
-	// If no meta flag is specified, check the default state of this warning flag
-	if (warningFlags[id].level == LEVEL_DEFAULT) { // enabled by default
-		return enabledBehavior;
-	}
-
-	// No flag enables this warning, explicitly or implicitly
-	return WarningBehavior::DISABLED;
-}
-
-void WarningState::update(WarningState other) {
-	if (other.state != WARNING_DEFAULT) {
-		state = other.state;
-	}
-	if (other.error != WARNING_DEFAULT) {
-		error = other.error;
-	}
-}
-
-void processWarningFlag(char const *flag) {
-	std::string rootFlag = flag;
-
-	// Check for `-Werror` or `-Wno-error` to return early
-	if (rootFlag == "error") {
-		// `-Werror` promotes warnings to errors
-		warningsAreErrors = true;
-		return;
-	} else if (rootFlag == "no-error") {
-		// `-Wno-error` disables promotion of warnings to errors
-		warningsAreErrors = false;
-		return;
-	}
-
-	// Check for prefixes that affect what the flag does
-	WarningState state;
-	if (rootFlag.starts_with("error=")) {
-		// `-Werror=<flag>` enables the flag as an error
-		state = {.state = WARNING_ENABLED, .error = WARNING_ENABLED};
-		rootFlag.erase(0, literal_strlen("error="));
-	} else if (rootFlag.starts_with("no-error=")) {
-		// `-Wno-error=<flag>` prevents the flag from being an error,
-		// without affecting whether it is enabled
-		state = {.state = WARNING_DEFAULT, .error = WARNING_DISABLED};
-		rootFlag.erase(0, literal_strlen("no-error="));
-	} else if (rootFlag.starts_with("no-")) {
-		// `-Wno-<flag>` disables the flag
-		state = {.state = WARNING_DISABLED, .error = WARNING_DEFAULT};
-		rootFlag.erase(0, literal_strlen("no-"));
-	} else {
-		// `-W<flag>` enables the flag
-		state = {.state = WARNING_ENABLED, .error = WARNING_DEFAULT};
-	}
-
-	// Check for an `=` parameter to process as a parametric warning
-	// `-Wno-<flag>` and `-Wno-error=<flag>` negation cannot have an `=` parameter, but without a
-	// parameter, the 0 value will apply to all levels of a parametric warning
-	uint8_t param = 0;
-	bool hasParam = false;
-	if (state.state == WARNING_ENABLED) {
-		// First, check if there is an "equals" sign followed by a decimal number
-		// Ignore an equal sign at the very end of the string
-		if (auto equals = rootFlag.find('=');
-		    equals != rootFlag.npos && equals != rootFlag.size() - 1) {
-			hasParam = true;
-
-			// Is the rest of the string a decimal number?
-			// We want to avoid `strtoul`'s whitespace and sign, so we parse manually
-			char const *ptr = rootFlag.c_str() + equals + 1;
-			bool warned = false;
-
-			// The `if`'s condition above ensures that this will run at least once
-			do {
-				// If we don't have a digit, bail
-				if (*ptr < '0' || *ptr > '9') {
-					break;
-				}
-				// Avoid overflowing!
-				if (param > UINT8_MAX - (*ptr - '0')) {
-					if (!warned) {
-						warnx("Invalid warning flag \"%s\": capping parameter at 255", flag);
-					}
-					warned = true; // Only warn once, cap always
-					param = 255;
-					continue;
-				}
-				param = param * 10 + (*ptr - '0');
-
-				ptr++;
-			} while (*ptr);
-
-			// If we reached the end of the string, truncate it at the '='
-			if (*ptr == '\0') {
-				rootFlag.resize(equals);
-				// `-W<flag>=0` is equivalent to `-Wno-<flag>`
-				if (param == 0) {
-					state.state = WARNING_DISABLED;
-				}
-			}
-		}
-	}
-
-	// Try to match the flag against a parametric warning
-	// If there was an equals sign, it will have set `param`; if not, `param` will be 0, which
-	// applies to all levels
-	for (auto const &paramWarning : paramWarnings) {
-		WarningID baseID = paramWarning.firstID;
-		uint8_t maxParam = paramWarning.lastID - baseID + 1;
-		assume(paramWarning.defaultLevel <= maxParam);
-
-		if (rootFlag == warningFlags[baseID].name) { // Match!
-			if (rootFlag == "numeric-string") {
-				warning(WARNING_OBSOLETE, "Warning flag \"numeric-string\" is deprecated\n");
-			}
-
-			// If making the warning an error but param is 0, set to the maximum
-			// This accommodates `-Werror=<flag>`, but also `-Werror=<flag>=0`, which is
-			// thus filtered out by the caller.
-			// A param of 0 makes sense for disabling everything, but neither for
-			// enabling nor "erroring". Use the default for those.
-			if (param == 0) {
-				param = paramWarning.defaultLevel;
-			} else if (param > maxParam) {
-				if (param != 255) { // Don't warn if already capped
-					warnx(
-					    "Invalid parameter %" PRIu8
-					    " for warning flag \"%s\"; capping at maximum %" PRIu8,
-					    param,
-					    rootFlag.c_str(),
-					    maxParam
-					);
-				}
-				param = maxParam;
-			}
-
-			// Set the first <param> to enabled/error, and disable the rest
-			for (uint8_t ofs = 0; ofs < maxParam; ofs++) {
-				WarningState &warning = warningStates.flagStates[baseID + ofs];
-				if (ofs < param) {
-					warning.update(state);
-				} else {
-					warning.state = WARNING_DISABLED;
-				}
-			}
-			return;
-		}
-	}
-
-	// Try to match against a non-parametric warning, unless there was an equals sign
-	if (!hasParam) {
-		// Try to match against a "meta" warning
-		for (WarningFlag const &metaWarning : metaWarnings) {
-			if (rootFlag == metaWarning.name) {
-				// Set each of the warning flags that meets this level
-				for (WarningID id : EnumSeq(NB_WARNINGS)) {
-					if (metaWarning.level >= warningFlags[id].level) {
-						warningStates.metaStates[id].update(state);
-					}
-				}
-				return;
-			}
-		}
-
-		// Try to match the flag against a "normal" flag
-		for (WarningID id : EnumSeq(NB_PLAIN_WARNINGS)) {
-			if (rootFlag == warningFlags[id].name) {
-				warningStates.flagStates[id].update(state);
-				return;
-			}
-		}
-	}
-
-	warnx("Unknown warning flag \"%s\"", flag);
-}
-
-void printDiag(
+static void printDiag(
     char const *fmt, va_list args, char const *type, char const *flagfmt, char const *flag
 ) {
 	fputs(type, stderr);
 	fputs(": ", stderr);
-	fstk_DumpCurrent();
+	if (fstk_DumpCurrent()) {
 		fprintf(stderr, flagfmt, flag);
 		fputs("\n    ", stderr);
+	}
 	vfprintf(stderr, fmt, args);
+	putc('\n', stderr);
+
 	lexer_DumpStringExpansions();
 }
 
+static void incrementErrors() {
+	// This intentionally makes 0 act as "unlimited"
+	warnings.incrementErrors();
+	if (warnings.nbErrors == options.maxErrors) {
+		fprintf(
+		    stderr,
+		    "Assembly aborted after the maximum of %" PRIu64 " error%s! (configure with "
+		    "'-X/--max-errors')\n",
+		    warnings.nbErrors,
+		    warnings.nbErrors == 1 ? "" : "s"
+		);
+		exit(1);
+	}
+}
+
 void error(char const *fmt, ...) {
 	va_list args;
 
@@ -314,20 +100,23 @@ void error(char const *fmt, ...) {
 	printDiag(fmt, args, "error", ":", nullptr);
 	va_end(args);
 
-	// This intentionally makes 0 act as "unlimited" (or at least "limited to sizeof(unsigned)")
-	nbErrors++;
-	if (nbErrors == maxErrors) {
-		errx(
-		    "The maximum of %u error%s was reached (configure with \"-X/--max-errors\"); assembly "
-		    "aborted!",
-		    maxErrors,
-		    maxErrors == 1 ? "" : "s"
-		);
+	incrementErrors();
+}
+
+void error(std::function<void()> callback) {
+	fputs("error: ", stderr);
+	if (fstk_DumpCurrent()) {
+		fputs(":\n    ", stderr);
 	}
+	callback();
+	putc('\n', stderr);
+	lexer_DumpStringExpansions();
+
+	incrementErrors();
 }
 
 [[noreturn]]
-void fatalerror(char const *fmt, ...) {
+void fatal(char const *fmt, ...) {
 	va_list args;
 
 	va_start(args, fmt);
@@ -337,13 +126,25 @@ void fatalerror(char const *fmt, ...) {
 	exit(1);
 }
 
+void requireZeroErrors() {
+	if (warnings.nbErrors != 0) {
+		fprintf(
+		    stderr,
+		    "Assembly aborted with %" PRIu64 " error%s!\n",
+		    warnings.nbErrors,
+		    warnings.nbErrors == 1 ? "" : "s"
+		);
+		exit(1);
+	}
+}
+
 void warning(WarningID id, char const *fmt, ...) {
-	char const *flag = warningFlags[id].name;
+	char const *flag = warnings.warningFlags[id].name;
 	va_list args;
 
 	va_start(args, fmt);
 
-	switch (getWarningBehavior(id)) {
+	switch (warnings.getWarningBehavior(id)) {
 	case WarningBehavior::DISABLED:
 		break;
 

src/diagnostics.cpp

@@ -0,0 +1,81 @@
+#include "diagnostics.hpp"
+
+void warnx(char const *fmt, ...) {
+	va_list ap;
+	fputs("warning: ", stderr);
+	va_start(ap, fmt);
+	vfprintf(stderr, fmt, ap);
+	va_end(ap);
+	putc('\n', stderr);
+}
+
+void WarningState::update(WarningState other) {
+	if (other.state != WARNING_DEFAULT) {
+		state = other.state;
+	}
+	if (other.error != WARNING_DEFAULT) {
+		error = other.error;
+	}
+}
+
+std::pair<WarningState, std::optional<uint32_t>> getInitialWarningState(std::string &flag) {
+	// Check for prefixes that affect what the flag does
+	WarningState state;
+	if (flag.starts_with("error=")) {
+		// `-Werror=<flag>` enables the flag as an error
+		state = {.state = WARNING_ENABLED, .error = WARNING_ENABLED};
+		flag.erase(0, literal_strlen("error="));
+	} else if (flag.starts_with("no-error=")) {
+		// `-Wno-error=<flag>` prevents the flag from being an error,
+		// without affecting whether it is enabled
+		state = {.state = WARNING_DEFAULT, .error = WARNING_DISABLED};
+		flag.erase(0, literal_strlen("no-error="));
+	} else if (flag.starts_with("no-")) {
+		// `-Wno-<flag>` disables the flag
+		state = {.state = WARNING_DISABLED, .error = WARNING_DEFAULT};
+		flag.erase(0, literal_strlen("no-"));
+	} else {
+		// `-W<flag>` enables the flag
+		state = {.state = WARNING_ENABLED, .error = WARNING_DEFAULT};
+	}
+
+	// Check if there is an "equals" sign followed by a decimal number
+	// Ignore an equals sign at the very end of the string
+	auto equals = flag.find('=');
+	// `-Wno-<flag>` and `-Wno-error=<flag>` negation cannot have an `=` parameter, but without
+	// one, the 0 value will apply to all levels of a parametric warning
+	if (state.state != WARNING_ENABLED || equals == flag.npos || equals == flag.size() - 1) {
+		return {state, std::nullopt};
+	}
+
+	// Is the rest of the string a decimal number?
+	// We want to avoid `strtoul`'s whitespace and sign, so we parse manually
+	char const *ptr = flag.c_str() + equals + 1;
+	uint32_t param = 0;
+	bool overflowed = false;
+
+	for (; *ptr >= '0' && *ptr <= '9'; ++ptr) {
+		if (overflowed) {
+			continue;
+		}
+
+		uint32_t c = *ptr - '0';
+		if (param > (UINT32_MAX - c) / 10) {
+			overflowed = true;
+			param = UINT32_MAX;
+			continue;
+		}
+		param = param * 10 + c;
+	}
+
+	// If we reached the end of the string, truncate it at the '='
+	if (*ptr == '\0') {
+		flag.resize(equals);
+		// `-W<flag>=0` is equivalent to `-Wno-<flag>`
+		if (param == 0) {
+			state.state = WARNING_DISABLED;
+		}
+	}
+
+	return {state, param};
+}

src/error.cpp

@@ -1,75 +0,0 @@
-// SPDX-License-Identifier: MIT
-
-#include "error.hpp"
-
-#include <errno.h>
-#include <stdarg.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
-static void vwarn(char const *fmt, va_list ap) {
-	char const *error = strerror(errno);
-
-	fprintf(stderr, "warning: ");
-	vfprintf(stderr, fmt, ap);
-	fprintf(stderr, ": %s\n", error);
-}
-
-static void vwarnx(char const *fmt, va_list ap) {
-	fprintf(stderr, "warning: ");
-	vfprintf(stderr, fmt, ap);
-	putc('\n', stderr);
-}
-
-[[noreturn]]
-static void verr(char const *fmt, va_list ap) {
-	char const *error = strerror(errno);
-
-	fprintf(stderr, "error: ");
-	vfprintf(stderr, fmt, ap);
-	fprintf(stderr, ": %s\n", error);
-	va_end(ap);
-	exit(1);
-}
-
-[[noreturn]]
-static void verrx(char const *fmt, va_list ap) {
-	fprintf(stderr, "error: ");
-	vfprintf(stderr, fmt, ap);
-	putc('\n', stderr);
-	va_end(ap);
-	exit(1);
-}
-
-void warn(char const *fmt, ...) {
-	va_list ap;
-
-	va_start(ap, fmt);
-	vwarn(fmt, ap);
-	va_end(ap);
-}
-
-void warnx(char const *fmt, ...) {
-	va_list ap;
-
-	va_start(ap, fmt);
-	vwarnx(fmt, ap);
-	va_end(ap);
-}
-
-[[noreturn]]
-void err(char const *fmt, ...) {
-	va_list ap;
-
-	va_start(ap, fmt);
-	verr(fmt, ap);
-}
-
-[[noreturn]]
-void errx(char const *fmt, ...) {
-	va_list ap;
-
-	va_start(ap, fmt);
-	verrx(fmt, ap);
-}

src/extern/getopt.cpp

@@ -1,6 +1,6 @@
 // SPDX-License-Identifier: MIT
 
-// This implementation was taken from musl and modified for RGBDS
+// This implementation was taken from musl and modified for RGBDS.
 
 #include "extern/getopt.hpp"
 
@@ -57,7 +57,7 @@ static int getopt(int argc, char *argv[], char const *optstring) {
 	}
 
 	if (!musl_optpos) {
-		musl_optpos++;
+		++musl_optpos;
 	}
 	k = mbtowc(&c, argv[musl_optind] + musl_optpos, MB_LEN_MAX);
 	if (k < 0) {
@@ -68,12 +68,12 @@ static int getopt(int argc, char *argv[], char const *optstring) {
 	musl_optpos += k;
 
 	if (!argv[musl_optind][musl_optpos]) {
-		musl_optind++;
+		++musl_optind;
 		musl_optpos = 0;
 	}
 
 	if (optstring[0] == '-' || optstring[0] == '+') {
-		optstring++;
+		++optstring;
 	}
 
 	i = 0;
@@ -83,7 +83,7 @@ static int getopt(int argc, char *argv[], char const *optstring) {
 		if (l > 0) {
 			i += l;
 		} else {
-			i++;
+			++i;
 		}
 	} while (l && d != c);
 
@@ -118,7 +118,7 @@ static void permute(char **argv, int dest, int src) {
 	char *tmp = argv[src];
 	int i;
 
-	for (i = src; i > dest; i--) {
+	for (i = src; i > dest; --i) {
 		argv[i] = argv[i - 1];
 	}
 	argv[dest] = tmp;
@@ -146,7 +146,7 @@ static int musl_getopt_long(
 	skipped = musl_optind;
 	if (optstring[0] != '+' && optstring[0] != '-') {
 		int i;
-		for (i = musl_optind;; i++) {
+		for (i = musl_optind;; ++i) {
 			if (i >= argc || !argv[i]) {
 				return -1;
 			}
@@ -161,7 +161,7 @@ static int musl_getopt_long(
 	if (resumed > skipped) {
 		int i, cnt = musl_optind - resumed;
 
-		for (i = 0; i < cnt; i++) {
+		for (i = 0; i < cnt; ++i) {
 			permute(argv, skipped, musl_optind - 1);
 		}
 		musl_optind = skipped + cnt;
@@ -180,16 +180,16 @@ static int musl_getopt_long_core(
 		int i, cnt, match = 0;
 		char *arg = 0, *opt, *start = argv[musl_optind] + 1;
 
-		for (cnt = i = 0; longopts[i].name; i++) {
+		for (cnt = i = 0; longopts[i].name; ++i) {
 			char const *name = longopts[i].name;
 
 			opt = start;
 			if (*opt == '-') {
-				opt++;
+				++opt;
 			}
 			while (*opt && *opt != '=' && *opt == *name) {
-				name++;
-				opt++;
+				++name;
+				++opt;
 			}
 			if (*opt && *opt != '=') {
 				continue;
@@ -200,19 +200,19 @@ static int musl_getopt_long_core(
 				cnt = 1;
 				break;
 			}
-			cnt++;
+			++cnt;
 		}
 		if (cnt == 1 && longonly && arg - start == mblen(start, MB_LEN_MAX)) {
 			int l = arg - start;
 
-			for (i = 0; optstring[i]; i++) {
+			for (i = 0; optstring[i]; ++i) {
 				int j = 0;
 
 				while (j < l && start[j] == optstring[i + j]) {
-					j++;
+					++j;
 				}
 				if (j == l) {
-					cnt++;
+					++cnt;
 					break;
 				}
 			}
@@ -220,7 +220,7 @@ static int musl_getopt_long_core(
 		if (cnt == 1) {
 			i = match;
 			opt = arg;
-			musl_optind++;
+			++musl_optind;
 			if (*opt == '=') {
 				if (!longopts[i].has_arg) {
 					musl_optopt = longopts[i].val;
@@ -254,7 +254,7 @@ static int musl_getopt_long_core(
 					);
 					return '?';
 				}
-				musl_optind++;
+				++musl_optind;
 			}
 			if (idx) {
 				*idx = i;
@@ -275,7 +275,7 @@ static int musl_getopt_long_core(
 				    strlen(argv[musl_optind] + 2)
 				);
 			}
-			musl_optind++;
+			++musl_optind;
 			return '?';
 		}
 	}

src/extern/utf8decoder.cpp

@@ -1,10 +1,15 @@
 // SPDX-License-Identifier: MIT
 
-// UTF-8 decoder: http://bjoern.hoehrmann.de/utf-8/decoder/dfa/
+// This implementation was taken from
+// http://bjoern.hoehrmann.de/utf-8/decoder/dfa/
+// and modified for RGBDS.
 
 #include "extern/utf8decoder.hpp"
 
+// clang-format off: vertically align values
 static uint8_t const utf8d[] = {
+    // The first part of the table maps bytes to character classes that
+    // to reduce the size of the transition table and create bitmasks.
      0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 00..0f
      0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 10..1f
      0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 20..2f
@@ -21,22 +26,23 @@ static uint8_t const utf8d[] = {
      2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // d0..df
     10, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 3, 3, // e0..ef
     11, 6, 6, 6, 5, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, // f0..ff
-    0,  1, 2, 3, 5, 8, 7, 1, 1, 1, 4, 6, 1, 1, 1, 1, // s0
-    1,  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // s1
-    1,  0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, // s1
-    1,  2, 1, 1, 1, 1, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, // s3
-    1,  1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, // s4
-    1,  2, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, // s5
-    1,  1, 1, 1, 1, 1, 1, 3, 1, 3, 1, 1, 1, 1, 1, 1, // s6
-    1,  3, 1, 1, 1, 1, 1, 3, 1, 3, 1, 1, 1, 1, 1, 1, // s7
-    1,  3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // s8
+    // The second part is a transition table that maps a combination
+    // of a state of the automaton and a character class to a state.
+     0, 12, 24, 36, 60, 96, 84, 12, 12, 12, 48, 72, // s0
+    12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, // s1
+    12,  0, 12, 12, 12, 12, 12,  0, 12,  0, 12, 12, // s2
+    12, 24, 12, 12, 12, 12, 12, 24, 12, 24, 12, 12, // s3
+    12, 12, 12, 12, 12, 12, 12, 24, 12, 12, 12, 12, // s4
+    12, 24, 12, 12, 12, 12, 12, 12, 12, 24, 12, 12, // s5
+    12, 12, 12, 12, 12, 12, 12, 36, 12, 36, 12, 12, // s6
+    12, 36, 12, 12, 12, 12, 12, 36, 12, 36, 12, 12, // s7
+    12, 36, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, // s8
 };
+// clang-format on
 
 uint32_t decode(uint32_t *state, uint32_t *codep, uint8_t byte) {
-	uint32_t type = utf8d[byte];
-
-	*codep = (*state != 0) ? (byte & 0x3FU) | (*codep << 6) : (0xFF >> type) & (byte);
-
-	*state = utf8d[256 + *state * 16 + type];
+	uint8_t type = utf8d[byte];
+	*codep = *state != UTF8_ACCEPT ? (byte & 0b111111) | (*codep << 6) : (0xff >> type) & byte;
+	*state = utf8d[0x100 + *state + type];
 	return *state;
 }

src/fix/mbc.cpp

@@ -0,0 +1,510 @@
+#include "fix/mbc.hpp"
+
+#include <stdlib.h>
+#include <unordered_map>
+#include <utility>
+
+#include "helpers.hpp"  // unreachable_
+#include "platform.hpp" // strcasecmp
+
+#include "fix/warning.hpp"
+
+// Associate every MBC type with its name and whether it has RAM
+static std::unordered_map<MbcType, std::pair<char const *, bool>> mbcData{
+    {ROM,                                   {"ROM", false}                           },
+    {ROM_RAM,                               {"ROM+RAM", true}                        },
+    {ROM_RAM_BATTERY,                       {"ROM+RAM+BATTERY", true}                },
+    {MBC1,                                  {"MBC1", false}                          },
+    {MBC1_RAM,                              {"MBC1+RAM", true}                       },
+    {MBC1_RAM_BATTERY,                      {"MBC1+RAM+BATTERY", true}               },
+    // MBC2 technically has RAM, but is not marked as such
+    {MBC2,                                  {"MBC2", false}                          },
+    {MBC2_BATTERY,                          {"MBC2+BATTERY", false}                  },
+    {MMM01,                                 {"MMM01", false}                         },
+    {MMM01_RAM,                             {"MMM01+RAM", true}                      },
+    {MMM01_RAM_BATTERY,                     {"MMM01+RAM+BATTERY", true}              },
+    {MBC3,                                  {"MBC3", false}                          },
+    {MBC3_TIMER_BATTERY,                    {"MBC3+TIMER+BATTERY", false}            },
+    {MBC3_TIMER_RAM_BATTERY,                {"MBC3+TIMER+RAM+BATTERY", true}         },
+    {MBC3_RAM,                              {"MBC3+RAM", true}                       },
+    {MBC3_RAM_BATTERY,                      {"MBC3+RAM+BATTERY", true}               },
+    {MBC5,                                  {"MBC5", false}                          },
+    {MBC5_RAM,                              {"MBC5+RAM", true}                       },
+    {MBC5_RAM_BATTERY,                      {"MBC5+RAM+BATTERY", true}               },
+    {MBC5_RUMBLE,                           {"MBC5+RUMBLE", false}                   },
+    {MBC5_RUMBLE_RAM,                       {"MBC5+RUMBLE+RAM", true}                },
+    {MBC5_RUMBLE_RAM_BATTERY,               {"MBC5+RUMBLE+RAM+BATTERY", true}        },
+    // MBC6 "Net de Get - Minigame @ 100" has RAM size 3 (32 KiB)
+    {MBC6,                                  {"MBC6", true}                           },
+    {MBC7_SENSOR_RUMBLE_RAM_BATTERY,        {"MBC7+SENSOR+RUMBLE+RAM+BATTERY", true} },
+    {POCKET_CAMERA,                         {"POCKET CAMERA", true}                  },
+    // Bandai TAMA5 "Game de Hakken!! Tamagotchi - Osutchi to Mesutchi" has RAM size 0
+    {BANDAI_TAMA5,                          {"BANDAI TAMA5", false}                  },
+    {HUC3,                                  {"HUC3", true}                           },
+    {HUC1_RAM_BATTERY,                      {"HUC1+RAM+BATTERY", true}               },
+    // TPP1 may or may not have RAM, don't use these flags for it
+    {TPP1,                                  {"TPP1", false}                          },
+    {TPP1_RUMBLE,                           {"TPP1+RUMBLE", false}                   },
+    {TPP1_MULTIRUMBLE_RUMBLE,               {"TPP1+MULTIRUMBLE", false}              },
+    {TPP1_TIMER,                            {"TPP1+TIMER", false}                    },
+    {TPP1_TIMER_RUMBLE,                     {"TPP1+TIMER+RUMBLE", false}             },
+    {TPP1_TIMER_MULTIRUMBLE_RUMBLE,         {"TPP1+TIMER+MULTIRUMBLE", false}        },
+    {TPP1_BATTERY,                          {"TPP1+BATTERY", false}                  },
+    {TPP1_BATTERY_RUMBLE,                   {"TPP1+BATTERY+RUMBLE", false}           },
+    {TPP1_BATTERY_MULTIRUMBLE_RUMBLE,       {"TPP1+BATTERY+MULTIRUMBLE", false}      },
+    {TPP1_BATTERY_TIMER,                    {"TPP1+BATTERY+TIMER", false}            },
+    {TPP1_BATTERY_TIMER_RUMBLE,             {"TPP1+BATTERY+TIMER+RUMBLE", false}     },
+    {TPP1_BATTERY_TIMER_MULTIRUMBLE_RUMBLE, {"TPP1+BATTERY+TIMER+MULTIRUMBLE", false}},
+};
+
+static char const *acceptedMBCNames =
+    "Accepted MBC names:\n"
+    "\tROM ($00) [aka ROM_ONLY]\n"
+    "\tMBC1 ($01), MBC1+RAM ($02), MBC1+RAM+BATTERY ($03)\n"
+    "\tMBC2 ($05), MBC2+BATTERY ($06)\n"
+    "\tROM+RAM ($08) [deprecated], ROM+RAM+BATTERY ($09) [deprecated]\n"
+    "\tMMM01 ($0B), MMM01+RAM ($0C), MMM01+RAM+BATTERY ($0D)\n"
+    "\tMBC3+TIMER+BATTERY ($0F), MBC3+TIMER+RAM+BATTERY ($10)\n"
+    "\tMBC3 ($11), MBC3+RAM ($12), MBC3+RAM+BATTERY ($13)\n"
+    "\tMBC5 ($19), MBC5+RAM ($1A), MBC5+RAM+BATTERY ($1B)\n"
+    "\tMBC5+RUMBLE ($1C), MBC5+RUMBLE+RAM ($1D), MBC5+RUMBLE+RAM+BATTERY ($1E)\n"
+    "\tMBC6 ($20)\n"
+    "\tMBC7+SENSOR+RUMBLE+RAM+BATTERY ($22)\n"
+    "\tPOCKET_CAMERA ($FC)\n"
+    "\tBANDAI_TAMA5 ($FD) [aka TAMA5]\n"
+    "\tHUC3 ($FE)\n"
+    "\tHUC1+RAM+BATTERY ($FF)\n"
+    "\n"
+    "\tTPP1_1.0, TPP1_1.0+RUMBLE, TPP1_1.0+MULTIRUMBLE, TPP1_1.0+TIMER,\n"
+    "\tTPP1_1.0+TIMER+RUMBLE, TPP1_1.0+TIMER+MULTIRUMBLE, TPP1_1.0+BATTERY,\n"
+    "\tTPP1_1.0+BATTERY+RUMBLE, TPP1_1.0+BATTERY+MULTIRUMBLE,\n"
+    "\tTPP1_1.0+BATTERY+TIMER, TPP1_1.0+BATTERY+TIMER+RUMBLE,\n"
+    "\tTPP1_1.0+BATTERY+TIMER+MULTIRUMBLE"; // No trailing newline
+
+char const *mbc_Name(MbcType type) {
+	auto search = mbcData.find(type);
+	return search != mbcData.end() ? search->second.first : "(unknown)";
+}
+
+bool mbc_HasRAM(MbcType type) {
+	auto search = mbcData.find(type);
+	return search != mbcData.end() && search->second.second;
+}
+
+static void skipWhitespace(char const *&ptr) {
+	while (*ptr == ' ' || *ptr == '\t') {
+		++ptr;
+	}
+}
+
+static void skipMBCSpace(char const *&ptr) {
+	while (*ptr == ' ' || *ptr == '\t' || *ptr == '_') {
+		++ptr;
+	}
+}
+
+static char normalizeMBCChar(char c) {
+	if (c >= 'a' && c <= 'z') { // Uppercase for comparison with `mbc_Name`s
+		c = c - 'a' + 'A';
+	} else if (c == '_') { // Treat underscores as spaces
+		c = ' ';
+	}
+	return c;
+}
+
+static bool readMBCSlice(char const *&name, char const *expected) {
+	while (*expected) {
+		// If `name` is too short, the character will be '\0' and this will return `false`
+		if (normalizeMBCChar(*name++) != *expected++) {
+			return false;
+		}
+	}
+	return true;
+}
+
+[[noreturn]]
+static void fatalUnknownMBC(char const *fullName) {
+	fatal("Unknown MBC \"%s\"\n%s", fullName, acceptedMBCNames);
+}
+
+[[noreturn]]
+static void fatalWrongMBCFeatures(char const *fullName) {
+	fatal("Features incompatible with MBC (\"%s\")\n%s", fullName, acceptedMBCNames);
+}
+
+MbcType mbc_ParseName(char const *name, uint8_t &tpp1Major, uint8_t &tpp1Minor) {
+	char const *fullName = name;
+
+	if (!strcasecmp(name, "help") || !strcasecmp(name, "list")) {
+		puts(acceptedMBCNames); // Outputs to stdout and appends a newline
+		exit(0);
+	}
+
+	if ((name[0] >= '0' && name[0] <= '9') || name[0] == '$') {
+		int base = 0;
+
+		if (name[0] == '$') {
+			++name;
+			base = 16;
+		}
+		// Parse number, and return it as-is (unless it's too large)
+		char *endptr;
+		unsigned long mbc = strtoul(name, &endptr, base);
+
+		if (*endptr) {
+			fatalUnknownMBC(fullName);
+		}
+		if (mbc > 0xFF) {
+			fatal("Specified MBC ID out of range 0-255: %s", fullName);
+		}
+		return static_cast<MbcType>(mbc);
+	}
+
+	// Begin by reading the MBC type:
+	uint16_t mbc;
+	char const *ptr = name;
+
+	skipWhitespace(ptr); // Trim off leading whitespace
+
+#define tryReadSlice(expected) \
+	do { \
+		if (!readMBCSlice(ptr, expected)) { \
+			fatalUnknownMBC(fullName); \
+		} \
+	} while (0)
+
+	switch (*ptr++) {
+	case 'R': // ROM / ROM_ONLY
+	case 'r':
+		tryReadSlice("OM");
+		// Handle optional " ONLY"
+		skipMBCSpace(ptr);
+		if (*ptr == 'O' || *ptr == 'o') {
+			++ptr;
+			tryReadSlice("NLY");
+		}
+		mbc = ROM;
+		break;
+
+	case 'M': // MBC{1, 2, 3, 5, 6, 7} / MMM01
+	case 'm':
+		switch (*ptr++) {
+		case 'B':
+		case 'b':
+			switch (*ptr++) {
+			case 'C':
+			case 'c':
+				break;
+			default:
+				fatalUnknownMBC(fullName);
+			}
+			switch (*ptr++) {
+			case '1':
+				mbc = MBC1;
+				break;
+			case '2':
+				mbc = MBC2;
+				break;
+			case '3':
+				mbc = MBC3;
+				break;
+			case '5':
+				mbc = MBC5;
+				break;
+			case '6':
+				mbc = MBC6;
+				break;
+			case '7':
+				mbc = MBC7_SENSOR_RUMBLE_RAM_BATTERY;
+				break;
+			default:
+				fatalUnknownMBC(fullName);
+			}
+			break;
+		case 'M':
+		case 'm':
+			tryReadSlice("M01");
+			mbc = MMM01;
+			break;
+		default:
+			fatalUnknownMBC(fullName);
+		}
+		break;
+
+	case 'P': // POCKET_CAMERA
+	case 'p':
+		tryReadSlice("OCKET CAMERA");
+		mbc = POCKET_CAMERA;
+		break;
+
+	case 'B': // BANDAI_TAMA5
+	case 'b':
+		tryReadSlice("ANDAI TAMA5");
+		mbc = BANDAI_TAMA5;
+		break;
+
+	case 'T': // TAMA5 / TPP1
+	case 't':
+		switch (*ptr++) {
+		case 'A':
+			tryReadSlice("MA5");
+			mbc = BANDAI_TAMA5;
+			break;
+		case 'P': {
+			tryReadSlice("P1");
+			// Parse version
+			skipMBCSpace(ptr);
+			// Major
+			char *endptr;
+			unsigned long val = strtoul(ptr, &endptr, 10);
+
+			if (endptr == ptr) {
+				fatal("Failed to parse TPP1 major revision number");
+			}
+			ptr = endptr;
+			if (val != 1) {
+				fatal("RGBFIX only supports TPP1 version 1.0");
+			}
+			tpp1Major = val;
+			tryReadSlice(".");
+			// Minor
+			val = strtoul(ptr, &endptr, 10);
+			if (endptr == ptr) {
+				fatal("Failed to parse TPP1 minor revision number");
+			}
+			ptr = endptr;
+			if (val > 0xFF) {
+				fatal("TPP1 minor revision number must be 8-bit");
+			}
+			tpp1Minor = val;
+			mbc = TPP1;
+			break;
+		}
+		default:
+			fatalUnknownMBC(fullName);
+		}
+		break;
+
+	case 'H': // HuC{1, 3}
+	case 'h':
+		tryReadSlice("UC");
+		switch (*ptr++) {
+		case '1':
+			mbc = HUC1_RAM_BATTERY;
+			break;
+		case '3':
+			mbc = HUC3;
+			break;
+		default:
+			fatalUnknownMBC(fullName);
+		}
+		break;
+
+	default:
+		fatalUnknownMBC(fullName);
+	}
+
+	// Read "additional features"
+	uint8_t features = 0;
+	// clang-format off: vertically align values
+	static constexpr uint8_t RAM         = 1 << 7;
+	static constexpr uint8_t BATTERY     = 1 << 6;
+	static constexpr uint8_t TIMER       = 1 << 5;
+	static constexpr uint8_t RUMBLE      = 1 << 4;
+	static constexpr uint8_t SENSOR      = 1 << 3;
+	static constexpr uint8_t MULTIRUMBLE = 1 << 2;
+	// clang-format on
+
+	for (;;) {
+		skipWhitespace(ptr); // Trim off trailing whitespace
+
+		// If done, start processing "features"
+		if (!*ptr) {
+			break;
+		}
+		// We expect a '+' at this point
+		skipMBCSpace(ptr);
+		if (*ptr++ != '+') {
+			fatalUnknownMBC(fullName);
+		}
+		skipMBCSpace(ptr);
+
+		switch (*ptr++) {
+		case 'B': // BATTERY
+		case 'b':
+			tryReadSlice("ATTERY");
+			features |= BATTERY;
+			break;
+
+		case 'M':
+		case 'm':
+			tryReadSlice("ULTIRUMBLE");
+			features |= MULTIRUMBLE;
+			break;
+
+		case 'R': // RAM or RUMBLE
+		case 'r':
+			switch (*ptr++) {
+			case 'U':
+			case 'u':
+				tryReadSlice("MBLE");
+				features |= RUMBLE;
+				break;
+			case 'A':
+			case 'a':
+				tryReadSlice("M");
+				features |= RAM;
+				break;
+			default:
+				fatalUnknownMBC(fullName);
+			}
+			break;
+
+		case 'S': // SENSOR
+		case 's':
+			tryReadSlice("ENSOR");
+			features |= SENSOR;
+			break;
+
+		case 'T': // TIMER
+		case 't':
+			tryReadSlice("IMER");
+			features |= TIMER;
+			break;
+
+		default:
+			fatalUnknownMBC(fullName);
+		}
+	}
+#undef tryReadSlice
+
+	switch (mbc) {
+	case ROM:
+		if (!features) {
+			break;
+		}
+		mbc = ROM_RAM - 1;
+		static_assert(ROM_RAM + 1 == ROM_RAM_BATTERY, "Enum sanity check failed!");
+		static_assert(MBC1 + 1 == MBC1_RAM, "Enum sanity check failed!");
+		static_assert(MBC1 + 2 == MBC1_RAM_BATTERY, "Enum sanity check failed!");
+		static_assert(MMM01 + 1 == MMM01_RAM, "Enum sanity check failed!");
+		static_assert(MMM01 + 2 == MMM01_RAM_BATTERY, "Enum sanity check failed!");
+		[[fallthrough]];
+	case MBC1:
+	case MMM01:
+		if (features == RAM) {
+			++mbc;
+		} else if (features == (RAM | BATTERY)) {
+			mbc += 2;
+		} else if (features) {
+			fatalWrongMBCFeatures(fullName);
+		}
+		break;
+
+	case MBC2:
+		if (features == BATTERY) {
+			mbc = MBC2_BATTERY;
+		} else if (features) {
+			fatalWrongMBCFeatures(fullName);
+		}
+		break;
+
+	case MBC3:
+		// Handle timer, which also requires battery
+		if (features & TIMER) {
+			if (!(features & BATTERY)) {
+				warning(WARNING_MBC, "MBC3+TIMER implies BATTERY");
+			}
+			features &= ~(TIMER | BATTERY); // Reset those bits
+			mbc = MBC3_TIMER_BATTERY;
+			// RAM is handled below
+		}
+		static_assert(MBC3 + 1 == MBC3_RAM, "Enum sanity check failed!");
+		static_assert(MBC3 + 2 == MBC3_RAM_BATTERY, "Enum sanity check failed!");
+		static_assert(
+		    MBC3_TIMER_BATTERY + 1 == MBC3_TIMER_RAM_BATTERY, "Enum sanity check failed!"
+		);
+		if (features == RAM) {
+			++mbc;
+		} else if (features == (RAM | BATTERY)) {
+			mbc += 2;
+		} else if (features) {
+			fatalWrongMBCFeatures(fullName);
+		}
+		break;
+
+	case MBC5:
+		if (features & RUMBLE) {
+			features &= ~RUMBLE;
+			mbc = MBC5_RUMBLE;
+		}
+		static_assert(MBC5 + 1 == MBC5_RAM, "Enum sanity check failed!");
+		static_assert(MBC5 + 2 == MBC5_RAM_BATTERY, "Enum sanity check failed!");
+		static_assert(MBC5_RUMBLE + 1 == MBC5_RUMBLE_RAM, "Enum sanity check failed!");
+		static_assert(MBC5_RUMBLE + 2 == MBC5_RUMBLE_RAM_BATTERY, "Enum sanity check failed!");
+		if (features == RAM) {
+			++mbc;
+		} else if (features == (RAM | BATTERY)) {
+			mbc += 2;
+		} else if (features) {
+			fatalWrongMBCFeatures(fullName);
+		}
+		break;
+
+	case MBC6:
+	case POCKET_CAMERA:
+	case BANDAI_TAMA5:
+	case HUC3:
+		// No extra features accepted
+		if (features) {
+			fatalWrongMBCFeatures(fullName);
+		}
+		break;
+
+	case MBC7_SENSOR_RUMBLE_RAM_BATTERY:
+		if (features != (SENSOR | RUMBLE | RAM | BATTERY)) {
+			fatalWrongMBCFeatures(fullName);
+		}
+		break;
+
+	case HUC1_RAM_BATTERY:
+		if (features != (RAM | BATTERY)) { // HuC1 expects RAM+BATTERY
+			fatalWrongMBCFeatures(fullName);
+		}
+		break;
+
+	case TPP1:
+		if (features & RAM) {
+			warning(WARNING_MBC, "TPP1 requests RAM implicitly if given a non-zero RAM size");
+		}
+		if (features & BATTERY) {
+			mbc |= 0x08;
+		}
+		if (features & TIMER) {
+			mbc |= 0x04;
+		}
+		if (features & MULTIRUMBLE) {
+			mbc |= 0x03; // Also set the rumble flag
+		}
+		if (features & RUMBLE) {
+			mbc |= 0x01;
+		}
+		if (features & SENSOR) {
+			fatalWrongMBCFeatures(fullName);
+		}
+		// Multiple rumble speeds imply rumble
+		if (mbc & 0x01) {
+			assume(mbc & 0x02);
+		}
+		break;
+	}
+
+	skipWhitespace(ptr); // Trim off trailing whitespace
+
+	// If there is still something past the whitespace, error out
+	if (*ptr) {
+		fatalUnknownMBC(fullName);
+	}
+
+	return static_cast<MbcType>(mbc);
+}