gbdev/rgbds
1
0
Fork 0
mirror of https://github.com/gbdev/rgbds.git synced 2025-11-20 18:22:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

Move more RGBASM parser action code out of the Bison file

Browse Source
This commit is contained in:
Rangi42
2025-08-02 20:52:09 -04:00
parent 2eeb333be0
commit 7be7483571
3 changed files with 253 additions and 218 deletions
Download Patch File Download Diff File Expand all files Collapse all files

36
include/asm/actions.hpp
View File

@@ -13,30 +13,46 @@
#include "asm/output.hpp" // AssertionType
#include "asm/rpn.hpp" // RPNCommand
struct AlignmentSpec {
uint8_t alignment;
uint16_t alignOfs;
};
void act_If(int32_t condition);
void act_Elif(int32_t condition);
void act_Else();
void act_Endc();
AlignmentSpec act_Alignment(int32_t alignment, int32_t alignOfs);
void act_Assert(AssertionType type, Expression const &expr, std::string const &message);
void act_StaticAssert(AssertionType type, int32_t condition, std::string const &message);
std::optional<std::string> act_ReadFile(std::string const &name, uint32_t maxLen);
uint32_t act_CharToNum(std::string const &str);
uint32_t act_StringToNum(std::vector<int32_t> const &str);
int32_t act_CharVal(std::string const &str, int32_t negIdx);
uint8_t act_StringByte(std::string const &str, int32_t negIdx);
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);
std::string
act_StringSlice(std::string const &str, int32_t negStart, std::optional<int32_t> negStop);
std::string act_StringSub(std::string const &str, int32_t negPos, std::optional<uint32_t> optLen);
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);
std::string act_StringChar(std::string const &str, int32_t negIdx);
std::string act_CharSub(std::string const &str, int32_t negPos);
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
);
std::string act_SectionName(std::string const &symName);
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

265
src/asm/actions.cpp
View File

@@ -15,6 +15,107 @@
#include "asm/symbol.hpp"
#include "asm/warning.hpp"
void act_If(int32_t condition) {
lexer_IncIFDepth();
if (condition) {
lexer_RunIFBlock();
} else {
lexer_SetMode(LEXER_SKIP_TO_ELIF);
}
}
void act_Elif(int32_t condition) {
if (lexer_GetIFDepth() == 0) {
fatal("Found ELIF outside of an IF construct");
}
if (lexer_RanIFBlock()) {
if (lexer_ReachedELSEBlock()) {
fatal("Found ELIF after an ELSE block");
}
lexer_SetMode(LEXER_SKIP_TO_ENDC);
} else if (condition) {
lexer_RunIFBlock();
} else {
lexer_SetMode(LEXER_SKIP_TO_ELIF);
}
}
void act_Else() {
if (lexer_GetIFDepth() == 0) {
fatal("Found ELSE outside of an IF construct");
}
if (lexer_RanIFBlock()) {
if (lexer_ReachedELSEBlock()) {
fatal("Found ELSE after an ELSE block");
}
lexer_SetMode(LEXER_SKIP_TO_ENDC);
} else {
lexer_RunIFBlock();
lexer_ReachELSEBlock();
}
}
void act_Endc() {
lexer_DecIFDepth();
}
AlignmentSpec act_Alignment(int32_t alignment, int32_t alignOfs) {
AlignmentSpec spec = {0, 0};
if (alignment > 16) {
error("Alignment must be between 0 and 16, not %u", alignment);
} else if (alignOfs <= -(1 << alignment) || alignOfs >= 1 << alignment) {
error(
"The absolute alignment offset (%" PRIu32 ") must be less than alignment size (%d)",
static_cast<uint32_t>(alignOfs < 0 ? -alignOfs : alignOfs),
1 << alignment
);
} else {
spec.alignment = alignment;
spec.alignOfs = alignOfs < 0 ? (1 << alignment) + alignOfs : alignOfs;
}
return spec;
}
static void failAssert(AssertionType type, std::string const &message) {
switch (type) {
case ASSERT_FATAL:
if (message.empty()) {
fatal("Assertion failed");
} else {
fatal("Assertion failed: %s", message.c_str());
}
case ASSERT_ERROR:
if (message.empty()) {
error("Assertion failed");
} else {
error("Assertion failed: %s", message.c_str());
}
break;
case ASSERT_WARN:
if (message.empty()) {
warning(WARNING_ASSERT, "Assertion failed");
} else {
warning(WARNING_ASSERT, "Assertion failed: %s", message.c_str());
}
break;
}
}
void act_Assert(AssertionType type, Expression const &expr, std::string const &message) {
if (!expr.isKnown()) {
out_CreateAssert(type, expr, message, sect_GetOutputOffset());
} else if (expr.value() == 0) {
failAssert(type, message);
}
}
void act_StaticAssert(AssertionType type, int32_t condition, std::string const &message) {
if (!condition) {
failAssert(type, message);
}
}
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) {
@@ -53,6 +154,15 @@ std::optional<std::string> act_ReadFile(std::string const &name, uint32_t maxLen
return contents;
}
uint32_t act_CharToNum(std::string const &str) {
if (std::vector<int32_t> output = charmap_Convert(str); output.size() == 1) {
return static_cast<uint32_t>(output[0]);
} else {
error("Character literals must be a single charmap unit");
return 0;
}
}
uint32_t act_StringToNum(std::vector<int32_t> const &str) {
uint32_t length = str.size();
@@ -80,6 +190,63 @@ uint32_t act_StringToNum(std::vector<int32_t> const &str) {
return r;
}
static uint32_t 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);
}
static uint32_t 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);
}
int32_t act_CharVal(std::string const &str, int32_t negIdx) {
if (size_t len = charmap_CharSize(str); len != 0) {
uint32_t idx = adjustNegativeIndex(negIdx, len, "CHARVAL");
if (std::optional<int32_t> val = charmap_CharValue(str, idx); val.has_value()) {
return *val;
} else {
warning(
WARNING_BUILTIN_ARG,
"CHARVAL: Index %" PRIu32 " is past the end of the character mapping",
idx
);
return 0;
}
} else {
error("CHARVAL: No character mapping for \"%s\"", str.c_str());
return 0;
}
}
uint8_t act_StringByte(std::string const &str, int32_t negIdx) {
size_t len = str.length();
if (uint32_t idx = adjustNegativeIndex(negIdx, len, "STRBYTE"); idx < len) {
return static_cast<uint8_t>(str[idx]);
} else {
warning(
WARNING_BUILTIN_ARG, "STRBYTE: Index %" PRIu32 " is past the end of the string", idx
);
return 0;
}
}
static void errorInvalidUTF8Byte(uint8_t byte, char const *functionName) {
error("%s: Invalid UTF-8 byte 0x%02hhX", functionName, byte);
}
@@ -116,7 +283,12 @@ size_t act_StringLen(std::string const &str, bool printErrors) {
return len;
}
std::string act_StringSlice(std::string const &str, uint32_t start, uint32_t stop) {
std::string
act_StringSlice(std::string const &str, int32_t negStart, std::optional<int32_t> negStop) {
size_t adjustLen = act_StringLen(str, false);
uint32_t start = adjustNegativeIndex(negStart, adjustLen, "STRSLICE");
uint32_t stop = negStop ? adjustNegativeIndex(*negStop, adjustLen, "STRSLICE") : adjustLen;
size_t strLen = str.length();
size_t index = 0;
uint32_t state = UTF8_ACCEPT;
@@ -180,7 +352,11 @@ std::string act_StringSlice(std::string const &str, uint32_t start, uint32_t sto
return str.substr(startIndex, index - startIndex);
}
std::string act_StringSub(std::string const &str, uint32_t pos, uint32_t len) {
std::string act_StringSub(std::string const &str, int32_t negPos, std::optional<uint32_t> optLen) {
size_t adjustLen = act_StringLen(str, false);
uint32_t pos = adjustNegativePos(negPos, adjustLen, "STRSUB");
uint32_t len = optLen ? *optLen : pos > adjustLen ? 0 : adjustLen + 1 - pos;
size_t strLen = str.length();
size_t index = 0;
uint32_t state = UTF8_ACCEPT;
@@ -248,7 +424,10 @@ size_t act_CharLen(std::string const &str) {
return len;
}
std::string act_StringChar(std::string const &str, uint32_t idx) {
std::string act_StringChar(std::string const &str, int32_t negIdx) {
size_t adjustLen = act_CharLen(str);
uint32_t idx = adjustNegativeIndex(negIdx, adjustLen, "STRCHAR");
std::string_view view = str;
size_t charLen = 1;
@@ -269,7 +448,10 @@ std::string act_StringChar(std::string const &str, uint32_t idx) {
return std::string(start);
}
std::string act_CharSub(std::string const &str, uint32_t pos) {
std::string act_CharSub(std::string const &str, int32_t negPos) {
size_t adjustLen = act_CharLen(str);
uint32_t pos = adjustNegativePos(negPos, adjustLen, "CHARSUB");
std::string_view view = str;
size_t charLen = 1;
@@ -317,32 +499,6 @@ 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) {
// 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()) {
@@ -431,39 +587,30 @@ std::string act_StringFormat(
return str;
}
std::string act_SectionName(std::string const &symName) {
Symbol *sym = sym_FindScopedValidSymbol(symName);
if (!sym) {
if (sym_IsPurgedScoped(symName)) {
fatal("Unknown symbol \"%s\"; it was purged", symName.c_str());
} else {
fatal("Unknown symbol \"%s\"", symName.c_str());
}
}
Section const *section = sym->getSection();
if (!section) {
fatal("\"%s\" does not belong to any section", sym->name.c_str());
}
return section->name;
}
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();
int32_t 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;
}
}

170
src/asm/parser.y
View File

@@ -12,16 +12,12 @@
#include "linkdefs.hpp"
#include "asm/actions.hpp"
#include "asm/lexer.hpp"
#include "asm/macro.hpp"
#include "asm/rpn.hpp"
#include "asm/section.hpp"
struct AlignmentSpec {
uint8_t alignment;
uint16_t alignOfs;
};
struct ForArgs {
int32_t start;
int32_t stop;
@@ -47,7 +43,6 @@
#include "extern/utf8decoder.hpp"
#include "helpers.hpp"
#include "asm/actions.hpp"
#include "asm/charmap.hpp"
#include "asm/fixpoint.hpp"
#include "asm/fstack.hpp"
@@ -465,48 +460,19 @@ line_directive:
if:
POP_IF iconst NEWLINE {
lexer_IncIFDepth();
if ($2) {
lexer_RunIFBlock();
} else {
lexer_SetMode(LEXER_SKIP_TO_ELIF);
}
act_If($2);
}
;
elif:
POP_ELIF iconst NEWLINE {
if (lexer_GetIFDepth() == 0) {
fatal("Found ELIF outside of an IF construct");
}
if (lexer_RanIFBlock()) {
if (lexer_ReachedELSEBlock()) {
fatal("Found ELIF after an ELSE block");
}
lexer_SetMode(LEXER_SKIP_TO_ENDC);
} else if ($2) {
lexer_RunIFBlock();
} else {
lexer_SetMode(LEXER_SKIP_TO_ELIF);
}
act_Elif($2);
}
;
else:
POP_ELSE NEWLINE {
if (lexer_GetIFDepth() == 0) {
fatal("Found ELSE outside of an IF construct");
}
if (lexer_RanIFBlock()) {
if (lexer_ReachedELSEBlock()) {
fatal("Found ELSE after an ELSE block");
}
lexer_SetMode(LEXER_SKIP_TO_ENDC);
} else {
lexer_RunIFBlock();
lexer_ReachELSEBlock();
}
act_Else();
}
;
@@ -521,7 +487,7 @@ plain_directive:
endc:
POP_ENDC {
lexer_DecIFDepth();
act_Endc();
}
;
@@ -684,29 +650,10 @@ align:
align_spec:
uconst {
if ($1 > 16) {
::error("Alignment must be between 0 and 16, not %u", $1);
$$.alignment = $$.alignOfs = 0;
} else {
$$.alignment = $1;
$$.alignOfs = 0;
}
$$ = act_Alignment($1, 0);
}
| uconst COMMA iconst {
if ($1 > 16) {
::error("Alignment must be between 0 and 16, not %u", $1);
$$.alignment = $$.alignOfs = 0;
} else if ($3 <= -(1 << $1) || $3 >= 1 << $1) {
::error(
"The absolute alignment offset (%" PRIu32 ") must be less than alignment size (%d)",
static_cast<uint32_t>($3 < 0 ? -$3 : $3),
1 << $1
);
$$.alignment = $$.alignOfs = 0;
} else {
$$.alignment = $1;
$$.alignOfs = $3 < 0 ? (1 << $1) + $3 : $3;
}
$$ = act_Alignment($1, $3);
}
;
@@ -796,28 +743,16 @@ assert_type:
assert:
POP_ASSERT assert_type relocexpr {
if (!$3.isKnown()) {
out_CreateAssert($2, $3, "", sect_GetOutputOffset());
} else if ($3.value() == 0) {
act_FailAssert($2);
}
act_Assert($2, $3, "");
}
| POP_ASSERT assert_type relocexpr COMMA string {
if (!$3.isKnown()) {
out_CreateAssert($2, $3, $5, sect_GetOutputOffset());
} else if ($3.value() == 0) {
act_FailAssertMsg($2, $5);
}
act_Assert($2, $3, $5);
}
| POP_STATIC_ASSERT assert_type iconst {
if ($3 == 0) {
act_FailAssert($2);
}
act_StaticAssert($2, $3, "");
}
| POP_STATIC_ASSERT assert_type iconst COMMA string {
if ($3 == 0) {
act_FailAssertMsg($2, $5);
}
act_StaticAssert($2, $3, $5);
}
;
@@ -963,13 +898,11 @@ ds:
}
| POP_DS POP_ALIGN LBRACK align_spec RBRACK trailing_comma {
uint32_t n = sect_GetAlignBytes($4.alignment, $4.alignOfs);
sect_Skip(n, true);
sect_AlignPC($4.alignment, $4.alignOfs);
}
| POP_DS POP_ALIGN LBRACK align_spec RBRACK COMMA ds_args trailing_comma {
uint32_t n = sect_GetAlignBytes($4.alignment, $4.alignOfs);
sect_RelBytes(n, $7);
sect_AlignPC($4.alignment, $4.alignOfs);
}
@@ -1385,13 +1318,7 @@ relocexpr_no_str:
$$.makeNumber($1);
}
| CHARACTER {
std::vector<int32_t> output = charmap_Convert($1);
if (output.size() == 1) {
$$.makeNumber(static_cast<uint32_t>(output[0]));
} else {
::error("Character literals must be a single charmap unit");
$$.makeNumber(0);
}
$$.makeNumber(act_CharToNum($1));
}
| OP_LOGICNOT relocexpr %prec NEG {
$$.makeUnaryOp(RPN_LOGNOT, std::move($2));
@@ -1598,36 +1525,10 @@ relocexpr_no_str:
$$.makeNumber(charSize);
}
| OP_CHARVAL LPAREN string COMMA iconst RPAREN {
if (size_t len = charmap_CharSize($3); len != 0) {
uint32_t idx = act_AdjustNegativeIndex($5, len, "CHARVAL");
if (std::optional<int32_t> val = charmap_CharValue($3, idx); val.has_value()) {
$$.makeNumber(*val);
} else {
warning(
WARNING_BUILTIN_ARG,
"CHARVAL: Index %" PRIu32 " is past the end of the character mapping",
idx
);
$$.makeNumber(0);
}
} else {
::error("CHARVAL: No character mapping for \"%s\"", $3.c_str());
$$.makeNumber(0);
}
$$.makeNumber(act_CharVal($3, $5));
}
| OP_STRBYTE LPAREN string COMMA iconst RPAREN {
size_t len = $3.length();
uint32_t idx = act_AdjustNegativeIndex($5, len, "STRBYTE");
if (idx < len) {
$$.makeNumber(static_cast<uint8_t>($3[idx]));
} else {
warning(
WARNING_BUILTIN_ARG,
"STRBYTE: Index %" PRIu32 " is past the end of the string",
idx
);
$$.makeNumber(0);
}
$$.makeNumber(act_StringByte($3, $5));
}
| LPAREN relocexpr RPAREN {
$$ = std::move($2);
@@ -1685,35 +1586,22 @@ string_literal:
}
}
| OP_STRSLICE LPAREN string COMMA iconst COMMA iconst RPAREN {
size_t len = act_StringLen($3, false);
uint32_t start = act_AdjustNegativeIndex($5, len, "STRSLICE");
uint32_t stop = act_AdjustNegativeIndex($7, len, "STRSLICE");
$$ = act_StringSlice($3, start, stop);
$$ = act_StringSlice($3, $5, $7);
}
| OP_STRSLICE LPAREN string COMMA iconst RPAREN {
size_t len = act_StringLen($3, false);
uint32_t start = act_AdjustNegativeIndex($5, len, "STRSLICE");
$$ = act_StringSlice($3, start, len);
$$ = act_StringSlice($3, $5, std::nullopt);
}
| OP_STRSUB LPAREN string COMMA iconst COMMA uconst RPAREN {
size_t len = act_StringLen($3, false);
uint32_t pos = act_AdjustNegativePos($5, len, "STRSUB");
$$ = act_StringSub($3, pos, $7);
$$ = act_StringSub($3, $5, $7);
}
| OP_STRSUB LPAREN string COMMA iconst RPAREN {
size_t len = act_StringLen($3, false);
uint32_t pos = act_AdjustNegativePos($5, len, "STRSUB");
$$ = act_StringSub($3, pos, pos > len ? 0 : len + 1 - pos);
$$ = act_StringSub($3, $5, std::nullopt);
}
| OP_STRCHAR LPAREN string COMMA iconst RPAREN {
size_t len = act_CharLen($3);
uint32_t idx = act_AdjustNegativeIndex($5, len, "STRCHAR");
$$ = act_StringChar($3, idx);
$$ = act_StringChar($3, $5);
}
| OP_CHARSUB LPAREN string COMMA iconst RPAREN {
size_t len = act_CharLen($3);
uint32_t pos = act_AdjustNegativePos($5, len, "CHARSUB");
$$ = act_CharSub($3, pos);
$$ = act_CharSub($3, $5);
}
| OP_REVCHAR LPAREN charmap_args RPAREN {
bool unique;
@@ -1745,23 +1633,7 @@ string_literal:
$$ = act_StringFormat($3.format, $3.args);
}
| POP_SECTION LPAREN scoped_sym RPAREN {
Symbol *sym = sym_FindScopedValidSymbol($3);
if (!sym) {
if (sym_IsPurgedScoped($3)) {
fatal("Unknown symbol \"%s\"; it was purged", $3.c_str());
} else {
fatal("Unknown symbol \"%s\"", $3.c_str());
}
}
Section const *section = sym->getSection();
if (!section) {
fatal("\"%s\" does not belong to any section", sym->name.c_str());
}
// Section names are capped by rgbasm's maximum string length,
// so this currently can't overflow.
$$ = section->name;
$$ = act_SectionName($3);
}
;