Files
dusk/src/dusk/asset/loader/locale/assetlocaleloader.c
T

807 lines
20 KiB
C

/**
* Copyright (c) 2026 Dominic Masters
*
* This software is released under the MIT License.
* https://opensource.org/licenses/MIT
*/
#include "assetlocaleloader.h"
#include "util/memory.h"
#include "util/math.h"
#include "util/string.h"
#include "assert/assert.h"
errorret_t assetLocaleFileInit(
assetlocalefile_t *localeFile,
const char_t *path
) {
assertNotNull(localeFile, "Locale file cannot be NULL.");
assertNotNull(path, "Locale file path cannot be NULL.");
memoryZero(localeFile, sizeof(assetlocalefile_t));
// Init the asset file.
errorChain(assetFileInit(&localeFile->file, path, NULL, NULL));
// Open the file handle
errorChain(assetFileOpen(&localeFile->file));
// Get the blank key, this is basically the header info for po files
char_t buffer[1024];
errorChain(assetLocaleGetString(localeFile, "", 0, buffer, sizeof(buffer)));
errorChain(assetLocaleParseHeader(localeFile, buffer, sizeof(buffer)));
errorOk();
}
errorret_t assetLocaleFileDispose(assetlocalefile_t *localeFile) {
assertNotNull(localeFile, "Locale file cannot be NULL.");
errorChain(assetFileClose(&localeFile->file));
errorChain(assetFileDispose(&localeFile->file));
errorOk();
}
errorret_t assetLocaleParseHeader(
assetlocalefile_t *localeFile,
char_t *headerBuffer,
const size_t headerBufferSize
) {
assertNotNull(localeFile, "Locale file cannot be NULL.");
assertNotNull(headerBuffer, "Header buffer cannot be NULL.");
assertTrue(headerBufferSize > 0, "Header buffer size must be > 0.");
// Find "Plural-Forms: " line and parse out plural form info
char_t *pluralFormsLine = strstr(headerBuffer, "Plural-Forms:");
if(!pluralFormsLine) {
errorOk();
}
pluralFormsLine += strlen("Plural-Forms:");
// Expect nplurals
char_t *npluralsStr = strstr(pluralFormsLine, "nplurals=");
if(!npluralsStr) {
errorThrow("Failed to find nplurals in Plural-Forms header.");
}
npluralsStr += strlen("nplurals=");
localeFile->pluralStateCount = (uint8_t)atoi(npluralsStr);
if(localeFile->pluralStateCount == 0) {
errorThrow("nplurals must be greater than 0.");
}
if(localeFile->pluralStateCount > ASSET_LOCALE_FILE_PLURAL_FORM_COUNT) {
errorThrow(
"nplurals exceeds maximum supported plural forms: %d > %d",
localeFile->pluralStateCount,
ASSET_LOCALE_FILE_PLURAL_FORM_COUNT
);
}
// Expect plural=
char_t *pluralStr = strstr(pluralFormsLine, "plural=");
if(!pluralStr) {
errorThrow("Failed to find plural in Plural-Forms header.");
}
pluralStr += strlen("plural=");
// Expect ( [expressions] )
char_t *openParen = strchr(pluralStr, '(');
char_t *closeParen = strrchr(pluralStr, ')');
if(!openParen || !closeParen || closeParen < openParen) {
errorThrow("Failed to find plural expression in Plural-Forms header.");
}
// Parse:
// n [op] value ? index : n [op] value ? index : ... : final_index
char_t *ptr = openParen + 1;
uint8_t pluralIndex = 0;
uint8_t definedCount = 0;
while(1) {
while(*ptr == ' ') ptr++;
// Allow grouped subexpressions like:
// (n<7 ? 2 : 3)
// or
// (((3)))
uint8_t parenDepth = 0;
while(*ptr == '(') {
parenDepth++;
ptr++;
while(*ptr == ' ') ptr++;
}
// Final fallback: just an integer
if(*ptr != 'n') {
char_t *endPtr = NULL;
int32_t finalIndex = (int32_t)strtol(ptr, &endPtr, 10);
if(endPtr == ptr) {
errorThrow("Expected final plural index.");
}
ptr = endPtr;
while(*ptr == ' ') ptr++;
while(parenDepth > 0) {
if(*ptr != ')') {
errorThrow("Expected ')' after final plural index.");
}
ptr++;
parenDepth--;
while(*ptr == ' ') ptr++;
}
if(*ptr != ')') {
errorThrow("Expected ')' at end of plural expression.");
}
if(finalIndex < 0 || finalIndex >= localeFile->pluralStateCount) {
errorThrow(
"Final plural expression index out of bounds: %d (nplurals: %d)",
finalIndex,
localeFile->pluralStateCount
);
}
localeFile->pluralDefaultIndex = (uint8_t)finalIndex;
definedCount++;
break;
}
if(pluralIndex >= localeFile->pluralStateCount - 1) {
errorThrow(
"Too many plural conditions. Expected %d conditional clauses for nplurals=%d.",
localeFile->pluralStateCount - 1,
localeFile->pluralStateCount
);
}
ptr++; // skip 'n'
while(*ptr == ' ') ptr++;
// Determine operator
assetlocalepluraloperation_t op;
if(strncmp(ptr, "==", 2) == 0) {
op = ASSET_LOCALE_PLURAL_OP_EQUAL;
ptr += 2;
} else if(strncmp(ptr, "!=", 2) == 0) {
op = ASSET_LOCALE_PLURAL_OP_NOT_EQUAL;
ptr += 2;
} else if(strncmp(ptr, "<=", 2) == 0) {
op = ASSET_LOCALE_PLURAL_OP_LESS_EQUAL;
ptr += 2;
} else if(strncmp(ptr, ">=", 2) == 0) {
op = ASSET_LOCALE_PLURAL_OP_GREATER_EQUAL;
ptr += 2;
} else if(*ptr == '<') {
op = ASSET_LOCALE_PLURAL_OP_LESS;
ptr++;
} else if(*ptr == '>') {
op = ASSET_LOCALE_PLURAL_OP_GREATER;
ptr++;
} else {
errorThrow("Unsupported plural operator.");
}
while(*ptr == ' ') ptr++;
// Parse the comparitor value
char_t *endPtr = NULL;
int32_t value = (int32_t)strtol(ptr, &endPtr, 10);
if(endPtr == ptr) {
errorThrow("Expected value for plural expression.");
}
ptr = endPtr;
while(*ptr == ' ') ptr++;
// Parse ternary operator
if(*ptr != '?') {
errorThrow("Expected '?' after plural expression.");
}
ptr++;
while(*ptr == ' ') ptr++;
// Parse the indice
endPtr = NULL;
int32_t index = (int32_t)strtol(ptr, &endPtr, 10);
if(endPtr == ptr) {
errorThrow("Expected index for plural expression.");
}
ptr = endPtr;
if(index < 0 || index >= localeFile->pluralStateCount) {
errorThrow(
"Plural expression index out of bounds: %d (nplurals: %d)",
index,
localeFile->pluralStateCount
);
}
// Store plural expression.
localeFile->pluralIndices[pluralIndex] = (uint8_t)index;
localeFile->pluralOps[pluralIndex] = op;
localeFile->pluralValues[pluralIndex] = value;
pluralIndex++;
definedCount++;
while(*ptr == ' ') ptr++;
// Close any grouping parens that wrapped this conditional branch
while(parenDepth > 0) {
if(*ptr != ')') {
break;
}
ptr++;
parenDepth--;
while(*ptr == ' ') ptr++;
}
if(*ptr != ':') {
errorThrow("Expected ':' after plural expression.");
}
ptr++;
}
// Must define exactly nplurals outcomes:
// (nplurals - 1) conditional results + 1 final fallback result
if(
pluralIndex != localeFile->pluralStateCount - 1 ||
definedCount != localeFile->pluralStateCount
) {
errorThrow("Plural expression count does not match nplurals.");
}
errorOk();
}
uint8_t assetLocaleEvaluatePlural(
assetlocalefile_t *file,
const int32_t pluralCount
) {
assertNotNull(file, "Locale file cannot be NULL.");
assertTrue(pluralCount >= 0, "Plural count cannot be negative.");
for(uint8_t i = 0; i < file->pluralStateCount - 1; i++) {
int32_t value = file->pluralValues[i];
switch(file->pluralOps[i]) {
case ASSET_LOCALE_PLURAL_OP_EQUAL:
if(pluralCount == value) return file->pluralIndices[i];
break;
case ASSET_LOCALE_PLURAL_OP_NOT_EQUAL:
if(pluralCount != value) return file->pluralIndices[i];
break;
case ASSET_LOCALE_PLURAL_OP_LESS:
if(pluralCount < value) return file->pluralIndices[i];
break;
case ASSET_LOCALE_PLURAL_OP_LESS_EQUAL:
if(pluralCount <= value) return file->pluralIndices[i];
break;
case ASSET_LOCALE_PLURAL_OP_GREATER:
if(pluralCount > value) return file->pluralIndices[i];
break;
case ASSET_LOCALE_PLURAL_OP_GREATER_EQUAL:
if(pluralCount >= value) return file->pluralIndices[i];
break;
}
}
return file->pluralDefaultIndex;
}
errorret_t assetLocaleLineSkipBlanks(
assetfilelinereader_t *reader,
uint8_t *lineBuffer
) {
while(!reader->eof) {
// Skip blank lines
if(lineBuffer[0] == '\0') {
errorChain(assetFileLineReaderNext(reader));
continue;
}
// Skip comment lines
if(lineBuffer[0] == '#') {
errorChain(assetFileLineReaderNext(reader));
continue;
}
// Is line only spaces?
size_t lineLength = strlen((char_t *)lineBuffer);
size_t i;
bool_t onlySpaces = true;
for(i = 0; i < lineLength; i++) {
if(lineBuffer[i] != ' ') {
onlySpaces = false;
break;
}
}
if(onlySpaces) {
errorChain(assetFileLineReaderNext(reader));
continue;
}
break;
}
errorOk();
}
errorret_t assetLocaleLineUnbuffer(
assetfilelinereader_t *reader,
uint8_t *lineBuffer,
uint8_t *stringBuffer,
const size_t stringBufferSize
) {
stringBuffer[0] = '\0';
// At the point this funciton is called, we are looking for an opening quote.
char_t *start = strchr((char_t *)lineBuffer, '"');
if(!start) {
errorThrow("Expected open (0) \"");
}
char *end = strchr(start + 1, '"');
if(!end) {
errorThrow("Expected close (0) \"");
}
*end = '\0';
if(strlen(start) >= stringBufferSize) {
errorThrow("String buffer overflow");
}
memoryCopy(stringBuffer, start + 1, strlen(start));
// Now start buffering lines out
while(!reader->eof) {
errorChain(assetFileLineReaderNext(reader));
// Skip blank lines
errorChain(assetLocaleLineSkipBlanks(reader, lineBuffer));
// Skip starting spaces
char_t *ptr = (char_t *)lineBuffer;
while(*ptr == ' ') {
ptr++;
}
// Only consider lines starting with quote
if(*ptr != '"') {
break;
}
ptr++; // move past first quote
bool_t escaping = false;
char_t *dest = (char_t *)stringBuffer + strlen((char_t *)stringBuffer);
while(*ptr) {
if(escaping) {
// Handle escape sequences
switch(*ptr) {
case 'n': *dest++ = '\n'; break;
case 't': *dest++ = '\t'; break;
case '\\': *dest++ = '\\'; break;
case '"': *dest++ = '"'; break;
default:
errorThrow("Unknown escape sequence: \\%c", *ptr);
}
escaping = false;
} else if(*ptr == '\\') {
escaping = true;
} else if(*ptr == '"') {
// End of string
break;
} else {
// Regular character
*dest++ = *ptr;
}
if((size_t)(dest - (char_t *)stringBuffer) >= stringBufferSize) {
errorThrow("String buffer overflow");
}
ptr++;
}
*dest = '\0';
}
errorOk();
}
errorret_t assetLocaleGetString(
assetlocalefile_t *file,
const char_t *messageId,
const int32_t pluralCount,
char_t *stringBuffer,
const size_t stringBufferSize
) {
assertNotNull(file, "Asset file cannot be NULL.");
assertNotNull(messageId, "Message ID cannot be NULL.");
assertTrue(pluralCount >= 0, "Plural index cannot be negative.");
assertNotNull(stringBuffer, "String buffer cannot be NULL.");
assertTrue(stringBufferSize > 0, "String buffer size must be > 0");
assetfilelinereader_t reader;
bool_t msgidFound = false, msgidPluralFound = false, msgstrFound = false;
uint8_t msgidBuffer[256];
uint8_t msgidPluralBuffer[256];
uint8_t readBuffer[1024];
uint8_t lineBuffer[1024];
uint8_t pluralIndex = 0xFF;
msgidBuffer[0] = '\0';
msgidPluralBuffer[0] = '\0';
stringBuffer[0] = '\0';
// Rewind and start reading lines.
errorChain(assetFileRewind(&file->file));
assetFileLineReaderInit(
&reader,
&file->file,
readBuffer,
sizeof(readBuffer),
lineBuffer,
sizeof(lineBuffer)
);
// Skip blanks, comments, etc and start looking for msgid's
errorChain(assetLocaleLineSkipBlanks(&reader, lineBuffer));
while(!reader.eof) {
// Is this msgid?
if(memoryCompare(lineBuffer, "msgid", 5) != 0) {
errorChain(assetFileLineReaderNext(&reader));
msgidBuffer[0] = '\0';
continue;
}
// Unbuffer the msgid
assetLocaleLineUnbuffer(
&reader, lineBuffer, (uint8_t *)msgidBuffer, sizeof(msgidBuffer)
);
// Is this the needle?
if(memoryCompare(msgidBuffer, messageId, strlen(messageId)) != 0) {
continue;
}
msgidFound = true;
break;
}
if(!msgidFound) {
errorThrow("Failed to find message ID: %s", messageId);
}
// We are either going to see a msgstr or a msgid_plural
while(!reader.eof) {
errorChain(assetLocaleLineSkipBlanks(&reader, lineBuffer));
// Is msgid_plural?
if(
!msgidPluralFound &&
memoryCompare(lineBuffer, "msgid_plural", 12) == 0
) {
// Yes, start reading plural ID
assetLocaleLineUnbuffer(
&reader,
lineBuffer,
(uint8_t *)msgidPluralBuffer,
sizeof(msgidPluralBuffer)
);
msgidPluralFound = true;
// At this point we determine the plural index to use by running the
// plural formula
pluralIndex = assetLocaleEvaluatePlural(
file,
pluralCount
);
continue;
}
// Should be msgstr if not plural.
if(memoryCompare(lineBuffer, "msgstr", 6) != 0) {
errorThrow("Expected msgstr after msgid, found: %s", lineBuffer);
continue;
}
// If plural we need an index
if(msgidPluralFound) {
// Skip blank chars
char_t *ptr = (char_t *)lineBuffer + 6;
while(*ptr == ' ') {
ptr++;
}
if(*ptr != '[') {
errorThrow("Expected [ for plural form, found: %s", lineBuffer);
}
ptr++; // move past [
// Parse until ]
char *end = strchr(ptr, ']');
if(!end) {
errorThrow("Expected ] for plural form, found: %s", lineBuffer);
}
*end = '\0';
int32_t index = atoi(ptr);
if(index != pluralIndex) {
// Not the plural form we want, skip to the next useable line
while(!reader.eof) {
errorChain(assetFileLineReaderNext(&reader));
errorChain(assetLocaleLineSkipBlanks(&reader, lineBuffer));
if(
lineBuffer[0] == '\"' ||
lineBuffer[0] == '\0' ||
lineBuffer[0] == '#'
) continue;
break;
}
continue;
}
// Undo damage to line buffer from unbuffering.
*end = ']';
}
// Parse out msgstr
errorChain(assetLocaleLineUnbuffer(
&reader, lineBuffer, (uint8_t *)stringBuffer, stringBufferSize
));
msgstrFound = true;
break;
};
if(!msgstrFound) {
errorThrow("Failed to find msgstr for message ID: %s", messageId);
}
errorOk();
}
errorret_t assetLocaleGetStringWithVA(
assetlocalefile_t *file,
const char_t *messageId,
const int32_t pluralIndex,
char_t *buffer,
const size_t bufferSize,
...
) {
assertNotNull(file, "Asset file cannot be NULL.");
assertNotNull(messageId, "Message ID cannot be NULL.");
assertNotNull(buffer, "Buffer cannot be NULL.");
assertTrue(bufferSize > 0, "Buffer size must be > 0.");
assertTrue(pluralIndex >= 0, "Plural cannot be negative.");
char_t *tempBuffer;
tempBuffer = memoryAllocate(bufferSize);
errorret_t ret = assetLocaleGetString(
file,
messageId,
pluralIndex,
tempBuffer,
bufferSize
);
if(ret.code != ERROR_OK) {
memoryFree(tempBuffer);
return ret;
}
va_list args;
va_start(args, bufferSize);
int result = vsnprintf(buffer, bufferSize, tempBuffer, args);
va_end(args);
memoryFree(tempBuffer);
if(result < 0) {
errorThrow("Failed to format locale string for ID: %s", messageId);
}
return ret;
}
errorret_t assetLocaleGetStringWithArgs(
assetlocalefile_t *file,
const char_t *id,
const int32_t plural,
char_t *buffer,
const size_t bufferSize,
const assetlocalearg_t *args,
const size_t argCount
) {
assertNotNull(id, "Message ID cannot be NULL.");
assertNotNull(buffer, "Buffer cannot be NULL.");
assertTrue(bufferSize > 0, "Buffer size must be > 0.");
assertTrue(plural >= 0, "Plural cannot be negative.");
assertTrue(
argCount == 0 || args != NULL,
"Args cannot be NULL when argCount > 0."
);
char_t *format = memoryAllocate(bufferSize);
if(format == NULL) {
errorThrow("Failed to allocate format buffer.");
}
errorret_t ret = assetLocaleGetString(
file,
id,
plural,
format,
bufferSize
);
if(ret.code != ERROR_OK) {
memoryFree(format);
return ret;
}
size_t inIndex = 0;
size_t outIndex = 0;
size_t nextArg = 0;
buffer[0] = '\0';
while(format[inIndex] != '\0') {
if(format[inIndex] != '%') {
if(outIndex + 1 >= bufferSize) {
memoryFree(format);
errorThrow("Formatted locale string buffer overflow for ID: %s", id);
}
buffer[outIndex++] = format[inIndex++];
continue;
}
inIndex++;
/* Escaped percent */
if(format[inIndex] == '%') {
if(outIndex + 1 >= bufferSize) {
memoryFree(format);
errorThrow("Formatted locale string buffer overflow for ID: %s", id);
}
buffer[outIndex++] = '%';
inIndex++;
continue;
}
if(nextArg >= argCount) {
memoryFree(format);
errorThrow("Not enough locale arguments for ID: %s", id);
}
{
char_t specBuffer[32];
char_t valueBuffer[256];
size_t specLength = 0;
int written = 0;
char_t specifier;
specBuffer[specLength++] = '%';
/* Allow flags / width / precision */
while(format[inIndex] != '\0') {
char_t ch = format[inIndex];
if(
ch == '-' || ch == '+' || ch == ' ' || ch == '#' || ch == '0' ||
ch == '.' || (ch >= '0' && ch <= '9')
) {
if(specLength + 1 >= sizeof(specBuffer)) {
memoryFree(format);
errorThrow("Format specifier too long for ID: %s", id);
}
specBuffer[specLength++] = ch;
inIndex++;
continue;
}
break;
}
if(format[inIndex] == '\0') {
memoryFree(format);
errorThrow("Incomplete format specifier for ID: %s", id);
}
specifier = format[inIndex++];
if(specifier != 's' && specifier != 'd' && specifier != 'f') {
memoryFree(format);
errorThrow(
"Unsupported format specifier '%%%c' for ID: %s",
specifier,
id
);
}
specBuffer[specLength++] = specifier;
specBuffer[specLength] = '\0';
switch(specifier) {
case 's':
if(args[nextArg].type != ASSET_LOCALE_ARG_STRING) {
memoryFree(format);
errorThrow("Expected string locale argument for ID: %s", id);
}
written = snprintf(
valueBuffer,
sizeof(valueBuffer),
specBuffer,
args[nextArg].stringValue ? args[nextArg].stringValue : ""
);
break;
case 'd':
if(args[nextArg].type != ASSET_LOCALE_ARG_INT) {
memoryFree(format);
errorThrow("Expected int locale argument for ID: %s", id);
}
written = snprintf(
valueBuffer,
sizeof(valueBuffer),
specBuffer,
args[nextArg].intValue
);
break;
case 'f':
if(
args[nextArg].type != ASSET_LOCALE_ARG_FLOAT &&
args[nextArg].type != ASSET_LOCALE_ARG_INT
) {
memoryFree(format);
errorThrow("Expected float or int locale argument for ID: %s", id);
}
float_t floatValue = (
args[nextArg].type == ASSET_LOCALE_ARG_FLOAT ?
args[nextArg].floatValue :
(float_t)args[nextArg].intValue
);
written = snprintf(
valueBuffer,
sizeof(valueBuffer),
specBuffer,
floatValue
);
break;
}
nextArg++;
if(written < 0) {
memoryFree(format);
errorThrow("Failed to format locale string for ID: %s", id);
}
if(outIndex + (size_t)written >= bufferSize) {
memoryFree(format);
errorThrow("Formatted locale string buffer overflow for ID: %s", id);
}
memoryCopy(buffer + outIndex, valueBuffer, (size_t)written);
outIndex += (size_t)written;
}
}
buffer[outIndex] = '\0';
memoryFree(format);
errorOk();
}