Commit Graph
80 Commits
Author SHA1 Message Date
Rangi 85f96d671f Allow empty one-past-the-end sections (#2221) 2026-10-21 15:58:30 -04:00
Rangi 170b341315 Refactor unassignedSections into a flat list sorted by constraints (#2218)
Also removes the now-redundant `nbSectionsToAssign` counter
2026-10-17 14:12:58 -04:00
Rangi 9d6dab13d8 Sort linked sections by alignment before size (#2201)
Note that Libbet has no canonical hash and is sensitive to floating section placement
2026-10-11 03:40:15 +02:00
RangiandISSOtm b732f59472 Improve handling of platform-specific build options (#2162)
* Opt into POSIX functions with `_POSIX_C_SOURCE` before any system
  headers are `#include`d
* Allow user to override `make develop`'s `WARNFLAGS` and `CXXFLAGS`
* Use `make develop` on 32-bit Cygwin with sanitizers disabled
* Do not unnecessarily redefine `fseek` and `ftell`
* Disable a false-positive `-Wno-null-dereference` on 32-bit Cygwin

Co-authored-by: ISSOtm <[email protected]>
2026-09-26 16:31:25 -04:00
Eldred Habert 63a06f02c3 Fix constrained empty sections escaping their regions (#2137) 2026-09-21 15:54:10 -04:00
Eldred Habert 44328d522b Avoid using snprintf where GCC can check sizes (#2121)
Interestingly, Clang's `-Wformat-overflow` seems to only warn
about buffers so small they will *always* underflow,
whereas GCC tries to infer whether they can overflow *at all*.
(I'm guessing they lean towards false-negatives and false-positives resp.)

Anyway, calling `sprintf` here remains safe, since GCC (notably, via CI)
checks our work, and it simplifies the code ever so slightly while
also providing an (admittedly negligible) performance improvement.

Note that there may be more locations where we could use `sprintf`,
but a cursory glance at our other uses of `snprintf` didn't seem fruitful.
(One way to test is to set the target buffer size to 1, and see if a
warning pops up for such a trivially-wrong size. If not, you can be certain
the compiler won't be able to help you with a more realistic size.)
2026-09-19 03:15:07 +02:00
Rangi 142eae4558 Fix repeated "And N more" in overlay error message (#2118) 2026-09-18 17:35:30 -04:00
ISSOtm aab22852d3 Factor out free space manipulation 2026-09-18 21:19:37 +02:00
ISSOtm 0e41ace1e3 Refactor section placement algorithm
Massively simplified, by not trying to shoehorn differently-behaving
cases into the same one loop!

This may have introduced bugs (the test suite has caught three different
oversights, two of which via external projects, in fact!), but this also
makes the logic clearer and more streamlined, so that we are also less
likely to have any latent or future bugs.

In particular, previously, the first iteration of the loop could attempt
placement at an address not matching the section's constraints,
which made advancing to the next target address unnecessarily complicated
(https://github.com/gbdev/rgbds/pull/2064#discussion_r3985466596),
among other weirdness. The code ended up being defensive, and thus the
overall logic was murky.

I'm also expecting that this should provide a performance improvement due
to being essentially a form of loop-invariant code motion (and very likely
one a compiler couldn't have performed automatically), though I haven't
measured.
2026-09-18 21:19:37 +02:00
ISSOtm ce5c8ce8ac Respect zero-size sections' align offsets
Fixes #2095

This code pre-dates the introduction of alignment offsets,
so it was never updated to take it into account.

That said, some of the code that I have factored out is a little awkward because
it will be used in an upcoming refactoring of the whole assignment loop,
since that turns out to be more byzantine than it ought to be due to
trying to handle a too disparate set of cases in a single loop.
2026-09-18 21:19:37 +02:00
ISSOtm d43ef53745 Remove unreachable bank range check
The only case where it is not generated by our own code
(which we ought to be `assume`ing is correct) is when the bank
is used-specified, but that is caught by an earlier check.
2026-09-18 21:19:37 +02:00
ISSOtm ef709fade6 Refactor and fix sect packing's bank iteration
Moving both pieces of related code together, as it were.
This also ends up changing the behaviour of scrambling,
where sections could “leak” out of the scrambling's bank pool
and resume being placed normally.

We have discussed it offline, and decided that this was a bug.
The test suite has been updated accordingly, which even gives
the occasion to move this test from a special case to the main
generic loop!
2026-09-18 21:19:37 +02:00
ISSOtm 8685bcbfbf Refactor all scrambling-related dispatch into a single func
Yay for (poor man's) pattern matching! 🦀
2026-09-18 21:19:37 +02:00
ISSOtm 1ca4282e31 Move a function closer to its usage point
Trying to keep code grouped together well.
2026-09-18 21:19:37 +02:00
ISSOtm f234796428 Avoid allocating for our linkdef names (#2111)
This reverts a change introduced in fd78a9ae8, though it wasn't that commit's main point
so I'm feeling okay with undoing that.

This feels like an overkill change, using a static string is good enough for this
since we never modify this. I have considered using `string_view` instead, to have
the best of both worlds, but that's not NUL-terminated so our print functions
get a little grumpy.
2026-09-17 13:58:30 -04:00
Rangi b4988afc8a Fix RGBLINK ignoring alignment constraints when placing empty sections
Fixes #2095
2026-09-17 11:16:18 -04:00
Rangi aa657cd481 Make array buffers for snprintf just large enough for their uint32_t and uint16_t types
Note that `bank` can store 9 bytes (8 digits plus NUL terminator)
even though `section.bank` should never go above 16-bit (4 digits).
2026-09-17 10:12:05 -04:00
ISSOtm 9f175f7923 Use SectionTypeInfo struct member functions for section type info 2026-09-17 09:33:21 -04:00
Rangi a662cad7ce Fix infinite loop when an aligned HRAM section cannot be placed (#2064) 2026-09-11 05:21:26 +02:00
Rangi ba4a7f2e6a Explicitly assume what was verified earlier by readSection 2026-07-15 22:56:43 -04:00
Rangi b0948566bb Fail-safe treat 16-bit alignment as a fixed address even in RGBLINK
RGBASM does this itself, so should not output sections with 16-bit
alignment, but an invalid object file could otherwise cause
`getPlacement` to infinitely loop.
2026-07-15 22:46:57 -04:00
Rangi 5869e0dc60 Fix RGBLINK scrambling algorithm to work for SRAM
Since SRAM sections start from 0, we had been using a signed
`int8_t curScrambleSRAM` counter to allow 0 as a scrambled bank,
but this actually just placed all floating SRAM sections in bank 0.
2026-07-13 22:39:41 -04:00
Rangi 08354cf573 Explicitly encode assumption that an overlapping section exists 2026-07-13 17:47:38 -04:00
Rangi 41fe1d8f25 Correct error message for unconstrained sections with overlay (#1879) 2025-12-28 19:21:53 -05:00
Rangi 3631fab63c Fix bug where an object's invalid ROMX bank of 0 could break rgblink (#1868) 2025-12-05 11:21:00 -05:00
Rangi 131bb97ebc Fix some rgblink object file input bugs found via fuzzing with AFL++ (#1867)
- ID numbers (for fstack nodes, sections, symbols, patches, etc)
  might be too large for their associated collection
- Enum values might be invalid
- Bank values might be out of range for their section types
2025-12-04 20:49:16 -05:00
Rangi 5c2c893ced Refactor getSectionDescription in src/link/assign.cpp 2025-11-16 17:37:01 -05:00
Rangi42 df5162edca Use loops instead of tail calls and musttail
gcc 15.2.1 20250813 complains "address of automatic variable can
escape to `musttail` call" from `-Wmaybe-musttail-local-addr`,
and guaranteeing tail-call optimization cross-platform is more
trouble than it's worth.
2025-10-27 12:05:27 -04:00
Rangi 2873e0b8c8 Use musttail attribute to guarantee tail recursion (#1849) 2025-10-20 15:56:22 -04:00
Rangi 0c9920d4a6 Use C++ iterator for fragment/union "pieces" of RGBLINK sections 2025-10-05 15:07:25 -04:00
Rangi42 1fc9ba86c4 Some RGBLINK refactoring
- Consistently refer to `Section` fragments/unions as "pieces" (renaming `.nextu`)
- Remove `Symbol`'s `.label()` accessors (use `std::get<Label>`)
- Move some `Label`-related logic into `Symbol` methods
2025-09-05 16:34:51 -04:00
Rangi 3d155d5695 Some refactoring and cleanup (#1806)
* Use clang-tidy `misc-include-cleaner` for IWYU `#include` cleanup

* Use `std::optional<size_t>` instead of `ssize_t`

* Rename some functions in linkdefs.hpp

* Fix header order
2025-08-20 16:09:04 -04:00
Rangi 752b273aec Extend RGBASM and RGBLINK verbosity flags to have multiple levels like RGBGFX (#1772) 2025-08-02 17:10:10 -04:00
Rangi42 d16751f56a Avoid hard-coding a redundant "FATAL:" in RGBLINK 2025-07-27 20:07:53 -04:00
Rangi42 2ce4cdbff6 Reduce deep nesting some more, including larger refactors to assign.cpp 2025-07-22 19:38:49 -04:00
Rangi42 c83b87e0a0 Make all non-extern globals static 2025-07-21 20:14:09 -04:00
Rangi42 5f333d9753 More refactoring around extern variables 2025-07-21 19:22:10 -04:00
Rangi42 d652212857 Group extern RGBLINK variables in an Options struct, like RGBGFX 2025-07-21 18:12:40 -04:00
Rangi42 14f5e16ae8 Prefer pre-inc/dec unless post-inc/dec are necessary 2025-07-19 16:18:35 -04:00
Rangi 0c96234532 Use concrete types instead of auto when convenient and not redundant (#1757) 2025-07-17 14:59:51 -04:00
Rangi42 e7a05b1db8 Refactor for better formatting 2025-07-11 10:59:03 -04:00
Rangi 34cf959c9d Refactor to reduce nesting depth some more (#1740) 2025-07-09 22:46:40 -04:00
Rangi 5e43ece578 Remove errx and errors.hpp (#1737) 2025-07-09 11:04:23 -04:00
Rangi42 8c50839109 Miscellaneous updates 2025-07-08 17:45:13 -04:00
Rangi 35962dedc4 Refactor warnings and errors (#1728)
* Remove `err` and `warn`, keep `errx` and `warnx`, using them in RGBGFX too

* Separate RGBGFX and RGBLINK warnings/errors from main options

* Separate `report` function into `error` and `fatal` messages

* Implicit newlines for most RGBASM errors
2025-07-08 12:58:23 -04:00
Rangi 0150eb4bf3 Exclude more lines from test coverage (#1663)
These fall into a few categories:
- `_unreachable()`
- Verbose print messages
- Errors that should never practically occur (alloc/read/write failure,
  more than UINT32_MAX anonymous labels, etc)
2025-02-17 04:56:10 -05:00
Rangi b35e9d86fb Remove redundant @-style doc comment tags (#1641) 2025-01-29 19:56:28 -05:00
Rangi42 cae31005f8 Always use braces with InsertBraces: true in .clang-format 2025-01-27 20:12:12 -05:00
Rangi42 b8b60207f5 Use // line comments not /* block comments 2025-01-27 20:12:12 -05:00
Rangi42 a354af3d08 Reformat source files with clang-format 19.1.7 2025-01-27 20:12:12 -05:00