tron@2186: /* $Id$ */
tron@2186: 
truelight@0: #include "../stdafx.h"
tron@3015: #include "../macros.h"
truelight@4370: #include "../string.h"
peter1138@5108: #include "../table/control_codes.h"
rubidium@5838: #include "../helpers.hpp"
truelight@0: #include <stdio.h>
truelight@0: #include <string.h>
truelight@0: #include <stdlib.h>
truelight@0: #include <stdarg.h>
truelight@0: 
Darkvater@2482: #if (!defined(WIN32) && !defined(WIN64)) || defined(__CYGWIN__)
truelight@0: #include <unistd.h>
rubidium@5838: #include <sys/types.h>
rubidium@5838: #include <sys/stat.h>
truelight@0: #endif
truelight@0: 
truelight@4370: #if defined WIN32 || defined __WATCOMC__
truelight@4370: #include <direct.h>
truelight@4370: #endif /* WIN32 || __WATCOMC__ */
truelight@4370: 
darkvater@181: #ifdef __MORPHOS__
darkvater@181: #ifdef stderr
darkvater@181: #undef stderr
darkvater@181: #endif
darkvater@181: #define stderr stdout
truelight@4370: #endif /* __MORPHOS__ */
darkvater@181: 
truelight@0: /* Compiles a list of strings into a compiled string list */
truelight@0: 
truelight@0: typedef void (*ParseCmdProc)(char *buf, int value);
truelight@0: 
bjarni@6298: struct LanguagePackHeader {
truelight@0: 	uint32 ident;
rubidium@4344: 	uint32 version;     // 32-bits of auto generated version info which is basically a hash of strings.h
rubidium@4344: 	char name[32];      // the international name of this language
rubidium@4434: 	char own_name[32];  // the localized name of this language
rubidium@4344: 	char isocode[16];   // the ISO code for the language (not country code)
rubidium@4344: 	uint16 offsets[32]; // the offsets
rubidium@4344: 	byte plural_form;   // plural form index
rubidium@4344: 	byte pad[3];        // pad header to be a multiple of 4
bjarni@6298: };
truelight@0: 
bjarni@6298: struct CmdStruct {
truelight@0: 	const char *cmd;
truelight@0: 	ParseCmdProc proc;
truelight@0: 	long value;
ludde@2063: 	int8 consumes;
ludde@2087: 	byte flags;
bjarni@6298: };
truelight@0: 
ludde@2087: enum {
ludde@2087: 	C_DONTCOUNT = 1,
rubidium@4344: 	C_CASE      = 2,
ludde@2087: };
ludde@2087: 
ludde@2087: 
bjarni@6298: struct Case {
ludde@2087: 	int caseidx;
ludde@2087: 	char *string;
bjarni@6298: 	Case *next;
bjarni@6298: };
ludde@2087: 
Darkvater@2586: static bool _masterlang;
Darkvater@2585: static bool _translated;
tron@2459: static const char* _file = "(unknown file)";
tron@2059: static int _cur_line;
ludde@2063: static int _errors, _warnings;
ludde@2063: 
bjarni@6298: struct LangString {
rubidium@4344: 	char *name;            // Name of the string
rubidium@4344: 	char *english;         // English text
rubidium@4344: 	char *translated;      // Translated text
rubidium@4344: 	uint16 hash_next;      // next hash entry
ludde@2087: 	uint16 index;
rubidium@4344: 	int line;              // line of string in source-file
rubidium@4344: 	Case *english_case;    // cases for english
rubidium@4344: 	Case *translated_case; // cases for foreign
bjarni@6298: };
ludde@2087: 
ludde@2087: static LangString *_strings[65536];
ludde@2087: 
ludde@2063: 
ludde@2063: #define HASH_SIZE 32767
ludde@2063: static uint16 _hash_head[HASH_SIZE];
ludde@2063: 
ludde@2063: static byte _put_buf[4096];
ludde@2063: static int _put_pos;
ludde@2063: static int _next_string_id;
truelight@0: 
tron@2059: static uint32 _hash;
ludde@2063: static char _lang_name[32], _lang_ownname[32], _lang_isocode[16];
ludde@2082: static byte _lang_pluralform;
ludde@2087: #define MAX_NUM_GENDER 8
ludde@2087: static char _genders[MAX_NUM_GENDER][8];
ludde@2087: static int _numgenders;
truelight@0: 
ludde@2087: // contains the name of all cases.
ludde@2087: #define MAX_NUM_CASES 50
ludde@2087: static char _cases[MAX_NUM_CASES][16];
ludde@2087: static int _numcases;
ludde@2084: 
ludde@2082: // for each plural value, this is the number of plural forms.
rubidium@4344: static const byte _plural_form_counts[] = { 2, 1, 2, 3, 3, 3, 3, 3, 4 };
ludde@2082: 
ludde@2082: static const char *_cur_ident;
tron@2059: 
bjarni@6298: struct CmdPair {
ludde@2087: 	const CmdStruct *a;
rubidium@5838: 	const char *v;
bjarni@6298: };
ludde@2087: 
bjarni@6298: struct ParsedCommandStruct {
ludde@2087: 	int np;
ludde@2087: 	CmdPair pairs[32];
ludde@2087: 	const CmdStruct *cmd[32]; // ordered by param #
bjarni@6298: };
ludde@2087: 
ludde@2087: // Used when generating some advanced commands.
ludde@2087: static ParsedCommandStruct _cur_pcs;
ludde@2087: static int _cur_argidx;
ludde@2087: 
tron@2059: static uint HashStr(const char *s)
tron@2059: {
tron@2059: 	uint hash = 0;
tron@3015: 	for (; *s != '\0'; s++) hash = ROL(hash, 3) ^ *s;
truelight@0: 	return hash % HASH_SIZE;
truelight@0: }
tron@2059: 
ludde@2087: static void HashAdd(const char *s, LangString *ls)
tron@2059: {
tron@2059: 	uint hash = HashStr(s);
ludde@2087: 	ls->hash_next = _hash_head[hash];
ludde@2087: 	_hash_head[hash] = ls->index + 1;
truelight@0: }
truelight@0: 
ludde@2087: static LangString *HashFind(const char *s)
tron@2059: {
ludde@2063: 	int idx = _hash_head[HashStr(s)];
tron@3015: 
ludde@2063: 	while (--idx >= 0) {
tron@3015: 		LangString* ls = _strings[idx];
tron@3015: 
tron@3015: 		if (strcmp(ls->name, s) == 0) return ls;
ludde@2087: 		idx = ls->hash_next;
truelight@0: 	}
ludde@2087: 	return NULL;
truelight@0: }
truelight@0: 
KUDr@3900: #ifdef _MSC_VER
KUDr@3900: # define LINE_NUM_FMT "(%d)"
KUDr@3900: #else
KUDr@3900: # define LINE_NUM_FMT ":%d"
KUDr@3900: #endif
ludde@2063: 
Darkvater@4922: static void CDECL warning(const char *s, ...)
tron@2059: {
truelight@0: 	char buf[1024];
truelight@0: 	va_list va;
truelight@0: 	va_start(va, s);
truelight@4370: 	vsnprintf(buf, lengthof(buf), s, va);
truelight@0: 	va_end(va);
Darkvater@4447: 	fprintf(stderr, "%s" LINE_NUM_FMT ": warning: %s\n", _file, _cur_line, buf);
ludde@2063: 	_warnings++;
truelight@0: }
truelight@0: 
Darkvater@4922: void CDECL error(const char *s, ...)
ludde@2063: {
ludde@2063: 	char buf[1024];
ludde@2063: 	va_list va;
ludde@2063: 	va_start(va, s);
truelight@4370: 	vsnprintf(buf, lengthof(buf), s, va);
ludde@2063: 	va_end(va);
Darkvater@4447: 	fprintf(stderr, "%s" LINE_NUM_FMT ": error: %s\n", _file, _cur_line, buf);
ludde@2063: 	_errors++;
ludde@2063: }
ludde@2063: 
ludde@2063: 
Darkvater@4922: static void NORETURN CDECL fatal(const char *s, ...)
tron@2059: {
truelight@0: 	char buf[1024];
truelight@0: 	va_list va;
truelight@0: 	va_start(va, s);
truelight@4370: 	vsnprintf(buf, lengthof(buf), s, va);
truelight@0: 	va_end(va);
KUDr@3900: 	fprintf(stderr, "%s" LINE_NUM_FMT ": FATAL: %s\n", _file, _cur_line, buf);
truelight@0: 	exit(1);
truelight@0: }
truelight@0: 
ludde@2063: static void PutByte(byte c)
tron@2059: {
Darkvater@4922: 	if (_put_pos == lengthof(_put_buf)) fatal("Put buffer too small");
truelight@0: 	_put_buf[_put_pos++] = c;
truelight@0: }
truelight@0: 
truelight@0: 
peter1138@5108: static void PutUtf8(uint32 value)
tron@2059: {
peter1138@5108: 	if (value < 0x80) {
peter1138@5108: 		PutByte(value);
peter1138@5108: 	} else if (value < 0x800) {
peter1138@5108: 		PutByte(0xC0 + GB(value,  6, 5));
peter1138@5108: 		PutByte(0x80 + GB(value,  0, 6));
peter1138@5108: 	} else if (value < 0x10000) {
peter1138@5108: 		PutByte(0xE0 + GB(value, 12, 4));
peter1138@5108: 		PutByte(0x80 + GB(value,  6, 6));
peter1138@5108: 		PutByte(0x80 + GB(value,  0, 6));
peter1138@5108: 	} else if (value < 0x110000) {
peter1138@5108: 		PutByte(0xF0 + GB(value, 18, 3));
peter1138@5108: 		PutByte(0x80 + GB(value, 12, 6));
peter1138@5108: 		PutByte(0x80 + GB(value,  6, 6));
peter1138@5108: 		PutByte(0x80 + GB(value,  0, 6));
peter1138@5108: 	} else {
Darkvater@5548: 		warning("Invalid unicode value U+0x%X", value);
peter1138@5108: 	}
truelight@0: }
truelight@0: 
truelight@0: 
peter1138@5111: size_t Utf8Validate(const char *s)
peter1138@5111: {
peter1138@5111: 	uint32 c;
peter1138@5111: 
peter1138@5111: 	if (!HASBIT(s[0], 7)) {
peter1138@5111: 		/* 1 byte */
peter1138@5111: 		return 1;
peter1138@5111: 	} else if (GB(s[0], 5, 3) == 6 && IsUtf8Part(s[1])) {
peter1138@5111: 		/* 2 bytes */
peter1138@5111: 		c = GB(s[0], 0, 5) << 6 | GB(s[1], 0, 6);
peter1138@5111: 		if (c >= 0x80) return 2;
peter1138@5111: 	} else if (GB(s[0], 4, 4) == 14 && IsUtf8Part(s[1]) && IsUtf8Part(s[2])) {
peter1138@5111: 		/* 3 bytes */
peter1138@5111: 		c = GB(s[0], 0, 4) << 12 | GB(s[1], 0, 6) << 6 | GB(s[2], 0, 6);
peter1138@5111: 		if (c >= 0x800) return 3;
peter1138@5111: 	} else if (GB(s[0], 3, 5) == 30 && IsUtf8Part(s[1]) && IsUtf8Part(s[2]) && IsUtf8Part(s[3])) {
peter1138@5111: 		/* 4 bytes */
peter1138@5111: 		c = GB(s[0], 0, 3) << 18 | GB(s[1], 0, 6) << 12 | GB(s[2], 0, 6) << 6 | GB(s[3], 0, 6);
peter1138@5111: 		if (c >= 0x10000 && c <= 0x10FFFF) return 4;
peter1138@5111: 	}
peter1138@5111: 
peter1138@5111: 	return 0;
peter1138@5111: }
peter1138@5111: 
peter1138@5111: 
peter1138@5108: static void EmitSingleChar(char *buf, int value)
tron@2059: {
Darkvater@4922: 	if (*buf != '\0') warning("Ignoring trailing letters in command");
peter1138@5108: 	PutUtf8(value);
ludde@2063: }
truelight@0: 
peter1138@5108: 
tron@2059: static void EmitSetX(char *buf, int value)
tron@2059: {
truelight@0: 	char *err;
truelight@0: 	int x = strtol(buf, &err, 0);
Darkvater@4922: 	if (*err != 0) fatal("SetX param invalid");
peter1138@5108: 	PutUtf8(SCC_SETX);
ludde@2063: 	PutByte((byte)x);
ludde@2063: }
tron@2059: 
truelight@0: 
tron@2059: static void EmitSetXY(char *buf, int value)
tron@2059: {
truelight@0: 	char *err;
tron@3015: 	int x;
tron@3015: 	int y;
tron@2059: 
tron@2059: 	x = strtol(buf, &err, 0);
Darkvater@4922: 	if (*err != ' ') fatal("SetXY param invalid");
tron@3015: 	y = strtol(err + 1, &err, 0);
Darkvater@4922: 	if (*err != 0) fatal("SetXY param invalid");
truelight@0: 
peter1138@5108: 	PutUtf8(SCC_SETXY);
ludde@2063: 	PutByte((byte)x);
ludde@2063: 	PutByte((byte)y);
truelight@0: }
truelight@0: 
ludde@2082: // The plural specifier looks like
ludde@2082: // {NUM} {PLURAL -1 passenger passengers} then it picks either passenger/passengers depending on the count in NUM
ludde@2082: 
ludde@2082: // This is encoded like
ludde@2082: //  CommandByte <ARG#> <NUM> {Length of each string} {each string}
ludde@2082: 
ludde@2087: bool ParseRelNum(char **buf, int *value)
ludde@2082: {
tron@3015: 	const char* s = *buf;
tron@3015: 	char* end;
ludde@2082: 	bool rel = false;
ludde@2087: 	int v;
ludde@2087: 
ludde@2082: 	while (*s == ' ' || *s == '\t') s++;
tron@3015: 	if (*s == '+') {
tron@3015: 		rel = true;
tron@3015: 		s++;
tron@3015: 	}
ludde@2087: 	v = strtol(s, &end, 0);
ludde@2082: 	if (end == s) return false;
tron@3015: 	if (rel || v < 0) {
ludde@2087: 		*value += v;
tron@3015: 	} else {
ludde@2087: 		*value = v;
tron@3015: 	}
ludde@2082: 	*buf = end;
ludde@2082: 	return true;
ludde@2082: }
ludde@2082: 
ludde@2082: // Parse out the next word, or NULL
ludde@2082: char *ParseWord(char **buf)
ludde@2082: {
ludde@2082: 	char *s = *buf, *r;
tron@3015: 
ludde@2082: 	while (*s == ' ' || *s == '\t') s++;
tron@3015: 	if (*s == '\0') return NULL;
ludde@2082: 
ludde@2082: 	if (*s == '"') {
ludde@2082: 		r = ++s;
ludde@2082: 		// parse until next " or NUL
tron@2952: 		for (;;) {
tron@3015: 			if (*s == '\0') break;
ludde@2082: 			if (*s == '"') {
tron@3015: 				*s++ = '\0';
ludde@2082: 				break;
ludde@2082: 			}
ludde@2082: 			s++;
ludde@2082: 		}
ludde@2082: 	} else {
ludde@2082: 		// proceed until whitespace or NUL
ludde@2082: 		r = s;
tron@2952: 		for (;;) {
tron@3015: 			if (*s == '\0') break;
ludde@2082: 			if (*s == ' ' || *s == '\t') {
tron@3015: 				*s++ = '\0';
ludde@2082: 				break;
ludde@2082: 			}
ludde@2082: 			s++;
ludde@2082: 		}
ludde@2082: 	}
ludde@2082: 	*buf = s;
ludde@2082: 	return r;
ludde@2082: }
ludde@2082: 
ludde@2082: // Forward declaration
ludde@2087: static int TranslateArgumentIdx(int arg);
ludde@2082: 
tron@3015: static void EmitWordList(const char* const* words, uint nw)
ludde@2084: {
tron@3015: 	uint i;
tron@3015: 	uint j;
ludde@2084: 
ludde@2084: 	PutByte(nw);
tron@2952: 	for (i = 0; i < nw; i++) PutByte(strlen(words[i]));
tron@2952: 	for (i = 0; i < nw; i++) {
tron@3015: 		for (j = 0; words[i][j] != '\0'; j++) PutByte(words[i][j]);
ludde@2084: 	}
ludde@2084: }
ludde@2084: 
ludde@2082: static void EmitPlural(char *buf, int value)
ludde@2082: {
ludde@2087: 	int argidx = _cur_argidx;
tron@3015: 	const char* words[5];
ludde@2082: 	int nw = 0;
ludde@2082: 
ludde@2087: 	// Parse out the number, if one exists. Otherwise default to prev arg.
tron@3015: 	if (!ParseRelNum(&buf, &argidx)) argidx--;
ludde@2082: 
ludde@2082: 	// Parse each string
Darkvater@2566: 	for (nw = 0; nw < 5; nw++) {
ludde@2082: 		words[nw] = ParseWord(&buf);
tron@3015: 		if (words[nw] == NULL) break;
ludde@2082: 	}
ludde@2082: 
ludde@2082: 	if (nw == 0)
Darkvater@4922: 		fatal("%s: No plural words", _cur_ident);
ludde@2082: 
tron@2765: 	if (_plural_form_counts[_lang_pluralform] != nw) {
Darkvater@2585: 		if (_translated) {
Darkvater@4922: 			fatal("%s: Invalid number of plural forms. Expecting %d, found %d.", _cur_ident,
Darkvater@2566: 				_plural_form_counts[_lang_pluralform], nw);
Darkvater@2566: 		} else {
Darkvater@4922: 			warning("'%s' is untranslated. Tweaking english string to allow compilation for plural forms", _cur_ident);
Darkvater@2566: 			if (nw > _plural_form_counts[_lang_pluralform]) {
Darkvater@2566: 				nw = _plural_form_counts[_lang_pluralform];
Darkvater@2566: 			} else {
tron@2952: 				for (; nw < _plural_form_counts[_lang_pluralform]; nw++) {
Darkvater@2566: 					words[nw] = words[nw - 1];
Darkvater@2566: 				}
Darkvater@2566: 			}
Darkvater@2566: 		}
tron@2765: 	}
ludde@2082: 
peter1138@5108: 	PutUtf8(SCC_PLURAL_LIST);
ludde@2087: 	PutByte(TranslateArgumentIdx(argidx));
ludde@2084: 	EmitWordList(words, nw);
ludde@2084: }
ludde@2084: 
ludde@2084: 
ludde@2084: static void EmitGender(char *buf, int value)
ludde@2084: {
ludde@2087: 	int argidx = _cur_argidx;
tron@3015: 	uint nw;
ludde@2084: 
ludde@2084: 	if (buf[0] == '=') {
ludde@2084: 		buf++;
ludde@2084: 
ludde@2084: 		// This is a {G=DER} command
tron@2952: 		for (nw = 0; ; nw++) {
Darkvater@4922: 			if (nw >= 8) fatal("G argument '%s' invalid", buf);
tron@3015: 			if (strcmp(buf, _genders[nw]) == 0) break;
ludde@2084: 		}
ludde@2084: 		// now nw contains the gender index
peter1138@5108: 		PutUtf8(SCC_GENDER_INDEX);
ludde@2084: 		PutByte(nw);
tron@3015: 	} else {
tron@3015: 		const char* words[8];
ludde@2084: 
ludde@2084: 		// This is a {G 0 foo bar two} command.
ludde@2087: 		// If no relative number exists, default to +0
ludde@2087: 		if (!ParseRelNum(&buf, &argidx)) {}
ludde@2084: 
tron@2952: 		for (nw = 0; nw < 8; nw++) {
ludde@2084: 			words[nw] = ParseWord(&buf);
tron@3015: 			if (words[nw] == NULL) break;
ludde@2084: 		}
Darkvater@4922: 		if (nw != _numgenders) fatal("Bad # of arguments for gender command");
peter1138@5108: 		PutUtf8(SCC_GENDER_LIST);
ludde@2087: 		PutByte(TranslateArgumentIdx(argidx));
ludde@2084: 		EmitWordList(words, nw);
ludde@2082: 	}
ludde@2082: }
ludde@2082: 
ludde@2082: 
truelight@0: static const CmdStruct _cmd_structs[] = {
truelight@0: 	// Update position
peter1138@5108: 	{"SETX",  EmitSetX,  SCC_SETX,  0, 0},
peter1138@5108: 	{"SETXY", EmitSetXY, SCC_SETXY, 0, 0},
truelight@0: 
truelight@0: 	// Font size
peter1138@5108: 	{"TINYFONT", EmitSingleChar, SCC_TINYFONT, 0, 0},
peter1138@5108: 	{"BIGFONT",  EmitSingleChar, SCC_BIGFONT,  0, 0},
truelight@0: 
truelight@0: 	// Colors
peter1138@5108: 	{"BLUE",    EmitSingleChar, SCC_BLUE,    0, 0},
peter1138@5108: 	{"SILVER",  EmitSingleChar, SCC_SILVER,  0, 0},
peter1138@5108: 	{"GOLD",    EmitSingleChar, SCC_GOLD,    0, 0},
peter1138@5108: 	{"RED",     EmitSingleChar, SCC_RED,     0, 0},
peter1138@5108: 	{"PURPLE",  EmitSingleChar, SCC_PURPLE,  0, 0},
peter1138@5108: 	{"LTBROWN", EmitSingleChar, SCC_LTBROWN, 0, 0},
peter1138@5108: 	{"ORANGE",  EmitSingleChar, SCC_ORANGE,  0, 0},
peter1138@5108: 	{"GREEN",   EmitSingleChar, SCC_GREEN,   0, 0},
peter1138@5108: 	{"YELLOW",  EmitSingleChar, SCC_YELLOW,  0, 0},
peter1138@5108: 	{"DKGREEN", EmitSingleChar, SCC_DKGREEN, 0, 0},
peter1138@5108: 	{"CREAM",   EmitSingleChar, SCC_CREAM,   0, 0},
peter1138@5108: 	{"BROWN",   EmitSingleChar, SCC_BROWN,   0, 0},
peter1138@5108: 	{"WHITE",   EmitSingleChar, SCC_WHITE,   0, 0},
peter1138@5108: 	{"LTBLUE",  EmitSingleChar, SCC_LTBLUE,  0, 0},
peter1138@5108: 	{"GRAY",    EmitSingleChar, SCC_GRAY,    0, 0},
peter1138@5108: 	{"DKBLUE",  EmitSingleChar, SCC_DKBLUE,  0, 0},
peter1138@5108: 	{"BLACK",   EmitSingleChar, SCC_BLACK,   0, 0},
darkvater@222: 
peter1138@5108: 	{"CURRCOMPACT",   EmitSingleChar, SCC_CURRENCY_COMPACT,    1, 0}, // compact currency (32 bits)
peter1138@5108: 	{"REV",           EmitSingleChar, SCC_REVISION,            0, 0}, // openttd revision string
peter1138@5108: 	{"SHORTCARGO",    EmitSingleChar, SCC_CARGO_SHORT,         2, 0}, // short cargo description, only ### tons, or ### litres
peter1138@5108: 	{"CURRCOMPACT64", EmitSingleChar, SCC_CURRENCY_COMPACT_64, 2, 0}, // compact currency 64 bits
truelight@0: 
tron@3015: 	// These are special versions of {STRING1}
tron@3015: 	// The first string includes the second string.
peter1138@5108: 	{"COMPANY",    EmitSingleChar, SCC_STRING1, 1, 0},
peter1138@5108: 	{"PLAYERNAME", EmitSingleChar, SCC_STRING1, 1, 0},
peter1138@5108: 	{"VEHICLE",    EmitSingleChar, SCC_STRING1, 1, 0},
ludde@2063: 
peter1138@5108: 	{"STRING1", EmitSingleChar, SCC_STRING1, 1, C_CASE}, // included string that consumes ONE argument
peter1138@5108: 	{"STRING2", EmitSingleChar, SCC_STRING2, 2, C_CASE}, // included string that consumes TWO arguments
peter1138@5108: 	{"STRING3", EmitSingleChar, SCC_STRING3, 3, C_CASE}, // included string that consumes THREE arguments
peter1138@5108: 	{"STRING4", EmitSingleChar, SCC_STRING4, 4, C_CASE}, // included string that consumes FOUR arguments
peter1138@5108: 	{"STRING5", EmitSingleChar, SCC_STRING5, 5, C_CASE}, // included string that consumes FIVE arguments
peter1138@5108: 
peter1138@5108: 	{"STATIONFEATURES", EmitSingleChar, SCC_STATION_FEATURES, 1, 0}, // station features string, icons of the features
peter1138@5108: 	{"INDUSTRY",        EmitSingleChar, SCC_INDUSTRY_NAME,    1, 0}, // industry, takes an industry #
peter1138@5108: 	{"CARGO",           EmitSingleChar, SCC_CARGO,            2, 0},
peter1138@5108: 	{"POWER",           EmitSingleChar, SCC_POWER,            1, 0},
peter1138@5108: 	{"VOLUME",          EmitSingleChar, SCC_VOLUME,           1, 0},
peter1138@5108: 	{"VOLUME_S",        EmitSingleChar, SCC_VOLUME_SHORT,     1, 0},
peter1138@5108: 	{"WEIGHT",          EmitSingleChar, SCC_WEIGHT,           1, 0},
peter1138@5108: 	{"WEIGHT_S",        EmitSingleChar, SCC_WEIGHT_SHORT,     1, 0},
peter1138@5108: 	{"FORCE",           EmitSingleChar, SCC_FORCE,            1, 0},
peter1138@5108: 	{"VELOCITY",        EmitSingleChar, SCC_VELOCITY,         1, 0},
ludde@2063: 
rubidium@4434: 	{"P", EmitPlural, 0, 0, C_DONTCOUNT}, // plural specifier
rubidium@4434: 	{"G", EmitGender, 0, 0, C_DONTCOUNT}, // gender specifier
ludde@2082: 
peter1138@5108: 	{"DATE_TINY",  EmitSingleChar, SCC_DATE_TINY, 1, 0},
peter1138@5108: 	{"DATE_SHORT", EmitSingleChar, SCC_DATE_SHORT, 1, 0},
peter1138@5108: 	{"DATE_LONG",  EmitSingleChar, SCC_DATE_LONG, 1, 0},
ludde@2063: 
peter1138@5108: 	{"SKIP", EmitSingleChar, SCC_SKIP, 1, 0},
tron@3015: 
peter1138@5108: 	{"STRING", EmitSingleChar, SCC_STRING, 1, C_CASE},
ludde@2063: 
tron@2410: 	// Numbers
peter1138@5108: 	{"COMMA", EmitSingleChar, SCC_COMMA, 1, 0}, // Number with comma
peter1138@5108: 	{"NUM",   EmitSingleChar, SCC_NUM,   1, 0}, // Signed number
tron@2410: 
peter1138@5108: 	{"CURRENCY",   EmitSingleChar, SCC_CURRENCY,    1, 0},
peter1138@5108: 	{"CURRENCY64", EmitSingleChar, SCC_CURRENCY_64, 2, 0},
tron@2410: 
peter1138@5108: 	{"WAYPOINT", EmitSingleChar, SCC_WAYPOINT_NAME, 1, 0}, // waypoint name
peter1138@5108: 	{"STATION",  EmitSingleChar, SCC_STATION_NAME,  1, 0},
peter1138@5108: 	{"TOWN",     EmitSingleChar, SCC_TOWN_NAME,     1, 0},
tron@3015: 
ludde@2087: 	// 0x9D is used for the pseudo command SETCASE
ludde@2087: 	// 0x9E is used for case switching
ludde@2087: 
peter1138@5108: 	{"",               EmitSingleChar, '\n',               0, C_DONTCOUNT},
peter1138@5108: 	{"{",              EmitSingleChar, '{',                0, C_DONTCOUNT},
peter1138@5108: 	{"UPARROW",        EmitSingleChar, SCC_UPARROW,        0, 0},
peter1138@5108: 	{"SMALLUPARROW",   EmitSingleChar, SCC_SMALLUPARROW,   0, 0},
peter1138@5108: 	{"SMALLDOWNARROW", EmitSingleChar, SCC_SMALLDOWNARROW, 0, 0},
peter1138@5108: 	{"TRAIN",          EmitSingleChar, SCC_TRAIN,          0, 0},
peter1138@5108: 	{"LORRY",          EmitSingleChar, SCC_LORRY,          0, 0},
peter1138@5108: 	{"BUS",            EmitSingleChar, SCC_BUS,            0, 0},
peter1138@5108: 	{"PLANE",          EmitSingleChar, SCC_PLANE,          0, 0},
peter1138@5108: 	{"SHIP",           EmitSingleChar, SCC_SHIP,           0, 0},
peter1138@5108: 	{"NBSP",           EmitSingleChar, 0xA0,               0, C_DONTCOUNT},
peter1138@5108: 	{"CENT",           EmitSingleChar, 0xA2,               0, C_DONTCOUNT},
peter1138@5108: 	{"POUNDSIGN",      EmitSingleChar, 0xA3,               0, C_DONTCOUNT},
peter1138@5108: 	{"EURO",           EmitSingleChar, 0x20AC,             0, C_DONTCOUNT},
peter1138@5108: 	{"YENSIGN",        EmitSingleChar, 0xA5,               0, C_DONTCOUNT},
peter1138@5108: 	{"COPYRIGHT",      EmitSingleChar, 0xA9,               0, C_DONTCOUNT},
peter1138@5108: 	{"DOWNARROW",      EmitSingleChar, SCC_DOWNARROW,      0, C_DONTCOUNT},
peter1138@5108: 	{"CHECKMARK",      EmitSingleChar, SCC_CHECKMARK,      0, C_DONTCOUNT},
peter1138@5108: 	{"CROSS",          EmitSingleChar, SCC_CROSS,          0, C_DONTCOUNT},
peter1138@5108: 	{"REGISTERED",     EmitSingleChar, 0xAE,               0, C_DONTCOUNT},
peter1138@5108: 	{"RIGHTARROW",     EmitSingleChar, SCC_RIGHTARROW,     0, C_DONTCOUNT},
peter1138@5108: 	{"SMALLLEFTARROW", EmitSingleChar, SCC_LESSTHAN,       0, C_DONTCOUNT},
peter1138@5108: 	{"SMALLRIGHTARROW",EmitSingleChar, SCC_GREATERTHAN,    0, C_DONTCOUNT},
truelight@0: };
truelight@0: 
ludde@2063: 
tron@2059: static const CmdStruct *FindCmd(const char *s, int len)
tron@2059: {
tron@3015: 	const CmdStruct* cs;
tron@3015: 
tron@3015: 	for (cs = _cmd_structs; cs != endof(_cmd_structs); cs++) {
tron@3015: 		if (strncmp(cs->cmd, s, len) == 0 && cs->cmd[len] == '\0') return cs;
truelight@0: 	}
truelight@0: 	return NULL;
truelight@0: }
truelight@0: 
tron@3015: static uint ResolveCaseName(const char *str, uint len)
ludde@2087: {
tron@3015: 	uint i;
tron@3015: 
tron@3015: 	for (i = 0; i < MAX_NUM_CASES; i++) {
tron@3015: 		if (memcmp(_cases[i], str, len) == 0 && _cases[i][len] == 0) return i + 1;
tron@3015: 	}
Darkvater@4922: 	fatal("Invalid case-name '%s'", str);
ludde@2087: }
ludde@2087: 
truelight@0: 
ludde@2063: // returns NULL on eof
ludde@2063: // else returns command struct
ludde@2087: static const CmdStruct *ParseCommandString(const char **str, char *param, int *argno, int *casei)
truelight@0: {
ludde@2087: 	const char *s = *str, *start;
ludde@2063: 	const CmdStruct *cmd;
tron@2060: 	byte c;
truelight@0: 
ludde@2063: 	*argno = -1;
ludde@2087: 	*casei = -1;
ludde@2063: 
ludde@2063: 	// Scan to the next command, exit if there's no next command.
tron@2952: 	for (; *s != '{'; s++) {
tron@3015: 		if (*s == '\0') return NULL;
ludde@2063: 	}
ludde@2063: 	s++; // Skip past the {
ludde@2063: 
ludde@2063: 	if (*s >= '0' && *s <= '9') {
ludde@2063: 		char *end;
tron@3015: 
ludde@2063: 		*argno = strtoul(s, &end, 0);
Darkvater@4922: 		if (*end != ':') fatal("missing arg #");
ludde@2063: 		s = end + 1;
ludde@2063: 	}
truelight@0: 
truelight@0: 	// parse command name
truelight@0: 	start = s;
ludde@2087: 	do {
truelight@0: 		c = *s++;
ludde@2087: 	} while (c != '}' && c != ' ' && c != '=' && c != '.' && c != 0);
truelight@0: 
tron@2059: 	cmd = FindCmd(start, s - start - 1);
truelight@0: 	if (cmd == NULL) {
Darkvater@4922: 		error("Undefined command '%.*s'", s - start - 1, start);
truelight@0: 		return NULL;
truelight@0: 	}
truelight@0: 
ludde@2087: 	if (c == '.') {
ludde@2087: 		const char *casep = s;
ludde@2087: 
ludde@2087: 		if (!(cmd->flags & C_CASE))
Darkvater@4922: 			fatal("Command '%s' can't have a case", cmd->cmd);
ludde@2087: 
ludde@2087: 		do c = *s++; while (c != '}' && c != ' ' && c != '\0');
tron@3015: 		*casei = ResolveCaseName(casep, s - casep - 1);
ludde@2087: 	}
ludde@2087: 
ludde@2087: 	if (c == '\0') {
Darkvater@4922: 		error("Missing } from command '%s'", start);
ludde@2087: 		return NULL;
ludde@2087: 	}
ludde@2087: 
ludde@2087: 
ludde@2084: 	if (c != '}') {
ludde@2084: 		if (c == '=') s--;
truelight@0: 		// copy params
truelight@0: 		start = s;
tron@2952: 		for (;;) {
truelight@0: 			c = *s++;
truelight@0: 			if (c == '}') break;
ludde@2063: 			if (c == '\0') {
Darkvater@4922: 				error("Missing } from command '%s'", start);
truelight@0: 				return NULL;
truelight@0: 			}
Darkvater@4922: 			if (s - start == 250) fatal("param command too long");
truelight@0: 			*param++ = c;
truelight@0: 		}
truelight@0: 	}
tron@3015: 	*param = '\0';
truelight@0: 
truelight@0: 	*str = s;
truelight@0: 
truelight@0: 	return cmd;
truelight@0: }
truelight@0: 
truelight@0: 
ludde@2063: static void HandlePragma(char *str)
truelight@0: {
ludde@2063: 	if (!memcmp(str, "id ", 3)) {
truelight@0: 		_next_string_id = strtoul(str + 3, NULL, 0);
ludde@2063: 	} else if (!memcmp(str, "name ", 5)) {
truelight@0: 		ttd_strlcpy(_lang_name, str + 5, sizeof(_lang_name));
ludde@2063: 	} else if (!memcmp(str, "ownname ", 8)) {
truelight@0: 		ttd_strlcpy(_lang_ownname, str + 8, sizeof(_lang_ownname));
ludde@2063: 	} else if (!memcmp(str, "isocode ", 8)) {
miham@1376: 		ttd_strlcpy(_lang_isocode, str + 8, sizeof(_lang_isocode));
ludde@2082: 	} else if (!memcmp(str, "plural ", 7)) {
ludde@2082: 		_lang_pluralform = atoi(str + 7);
ludde@2082: 		if (_lang_pluralform >= lengthof(_plural_form_counts))
Darkvater@4922: 			fatal("Invalid pluralform %d", _lang_pluralform);
ludde@2084: 	} else if (!memcmp(str, "gender ", 7)) {
tron@3015: 		char* buf = str + 7;
tron@3015: 
tron@2952: 		for (;;) {
tron@3015: 			const char* s = ParseWord(&buf);
tron@3015: 
tron@3015: 			if (s == NULL) break;
Darkvater@4922: 			if (_numgenders >= MAX_NUM_GENDER) fatal("Too many genders, max %d", MAX_NUM_GENDER);
ludde@2087: 			ttd_strlcpy(_genders[_numgenders], s, sizeof(_genders[_numgenders]));
ludde@2084: 			_numgenders++;
ludde@2084: 		}
ludde@2087: 	} else if (!memcmp(str, "case ", 5)) {
tron@3015: 		char* buf = str + 5;
tron@3015: 
tron@2952: 		for (;;) {
tron@3015: 			const char* s = ParseWord(&buf);
tron@3015: 
tron@3015: 			if (s == NULL) break;
Darkvater@4922: 			if (_numcases >= MAX_NUM_CASES) fatal("Too many cases, max %d", MAX_NUM_CASES);
ludde@2087: 			ttd_strlcpy(_cases[_numcases], s, sizeof(_cases[_numcases]));
ludde@2087: 			_numcases++;
ludde@2087: 		}
truelight@0: 	} else {
Darkvater@4922: 		fatal("unknown pragma '%s'", str);
truelight@0: 	}
truelight@0: }
truelight@0: 
tron@3015: static void ExtractCommandString(ParsedCommandStruct* p, const char* s, bool warnings)
truelight@0: {
truelight@0: 	char param[100];
ludde@2063: 	int argno;
ludde@2063: 	int argidx = 0;
ludde@2087: 	int casei;
truelight@0: 
ludde@2063: 	memset(p, 0, sizeof(*p));
truelight@0: 
tron@2952: 	for (;;) {
ludde@2063: 		// read until next command from a.
tron@3015: 		const CmdStruct* ar = ParseCommandString(&s, param, &argno, &casei);
tron@3015: 
tron@3015: 		if (ar == NULL) break;
truelight@0: 
ludde@2063: 		// Sanity checking
Darkvater@4922: 		if (argno != -1 && ar->consumes == 0) fatal("Non consumer param can't have a paramindex");
ludde@2063: 
ludde@2063: 		if (ar->consumes) {
tron@3015: 			if (argno != -1) argidx = argno;
Darkvater@4922: 			if (argidx < 0 || argidx >= lengthof(p->cmd)) fatal("invalid param idx %d", argidx);
Darkvater@4922: 			if (p->cmd[argidx] != NULL && p->cmd[argidx] != ar) fatal("duplicate param idx %d", argidx);
ludde@2063: 
ludde@2063: 			p->cmd[argidx++] = ar;
ludde@2087: 		} else if (!(ar->flags & C_DONTCOUNT)) { // Ignore some of them
Darkvater@4922: 			if (p->np >= lengthof(p->pairs)) fatal("too many commands in string, max %d", lengthof(p->pairs));
ludde@2063: 			p->pairs[p->np].a = ar;
tron@3015: 			p->pairs[p->np].v = param[0] != '\0' ? strdup(param) : "";
ludde@2063: 			p->np++;
ludde@2063: 		}
ludde@2063: 	}
truelight@0: }
truelight@0: 
ludde@2063: 
ludde@2063: static const CmdStruct *TranslateCmdForCompare(const CmdStruct *a)
tron@2059: {
tron@3015: 	if (a == NULL) return NULL;
ludde@2063: 
tron@3015: 	if (strcmp(a->cmd, "STRING1") == 0 ||
tron@3015: 			strcmp(a->cmd, "STRING2") == 0 ||
tron@3015: 			strcmp(a->cmd, "STRING3") == 0 ||
tron@3015: 			strcmp(a->cmd, "STRING4") == 0 ||
tron@3015: 			strcmp(a->cmd, "STRING5") == 0) {
ludde@2063: 		return FindCmd("STRING", 6);
tron@3015: 	}
ludde@2063: 
tron@3015: 	if (strcmp(a->cmd, "SKIP") == 0) return NULL;
ludde@2063: 
ludde@2063: 	return a;
ludde@2063: }
ludde@2063: 
ludde@2063: 
ludde@2069: static bool CheckCommandsMatch(char *a, char *b, const char *name)
ludde@2063: {
ludde@2063: 	ParsedCommandStruct templ;
ludde@2063: 	ParsedCommandStruct lang;
KUDr@6308: 	int i, j;
ludde@2063: 	bool result = true;
ludde@2063: 
ludde@2063: 	ExtractCommandString(&templ, b, true);
ludde@2063: 	ExtractCommandString(&lang, a, true);
ludde@2063: 
ludde@2063: 	// For each string in templ, see if we find it in lang
ludde@2063: 	if (templ.np != lang.np) {
Darkvater@4922: 		warning("%s: template string and language string have a different # of commands", name);
ludde@2063: 		result = false;
ludde@2063: 	}
ludde@2063: 
tron@2952: 	for (i = 0; i < templ.np; i++) {
ludde@2063: 		// see if we find it in lang, and zero it out
ludde@2063: 		bool found = false;
tron@2952: 		for (j = 0; j < lang.np; j++) {
ludde@2063: 			if (templ.pairs[i].a == lang.pairs[j].a &&
tron@3015: 					strcmp(templ.pairs[i].v, lang.pairs[j].v) == 0) {
ludde@2063: 				// it was found in both. zero it out from lang so we don't find it again
ludde@2063: 				lang.pairs[j].a = NULL;
ludde@2063: 				found = true;
ludde@2063: 				break;
ludde@2063: 			}
ludde@2063: 		}
ludde@2063: 
ludde@2063: 		if (!found) {
Darkvater@4922: 			warning("%s: command '%s' exists in template file but not in language file", name, templ.pairs[i].a->cmd);
ludde@2063: 			result = false;
ludde@2063: 		}
ludde@2063: 	}
ludde@2063: 
ludde@2063: 	// if we reach here, all non consumer commands match up.
ludde@2063: 	// Check if the non consumer commands match up also.
tron@2952: 	for (i = 0; i < lengthof(templ.cmd); i++) {
ludde@2063: 		if (TranslateCmdForCompare(templ.cmd[i]) != TranslateCmdForCompare(lang.cmd[i])) {
Darkvater@4922: 			warning("%s: Param idx #%d '%s' doesn't match with template command '%s'", name, i,
tron@3015: 				lang.cmd[i]  == NULL ? "<empty>" : lang.cmd[i]->cmd,
tron@3015: 				templ.cmd[i] == NULL ? "<empty>" : templ.cmd[i]->cmd);
ludde@2063: 			result = false;
ludde@2063: 		}
ludde@2063: 	}
ludde@2063: 
ludde@2063: 	return result;
ludde@2063: }
ludde@2063: 
ludde@2063: static void HandleString(char *str, bool master)
ludde@2063: {
ludde@2063: 	char *s,*t;
ludde@2087: 	LangString *ent;
ludde@2087: 	char *casep;
darkvater@222: 
truelight@0: 	if (*str == '#') {
tron@3015: 		if (str[1] == '#' && str[2] != '#') HandlePragma(str + 2);
truelight@0: 		return;
truelight@0: 	}
truelight@0: 
truelight@0: 	// Ignore comments & blank lines
tron@3015: 	if (*str == ';' || *str == ' ' || *str == '\0') return;
truelight@0: 
truelight@0: 	s = strchr(str, ':');
truelight@0: 	if (s == NULL) {
Darkvater@4922: 		error("Line has no ':' delimiter");
truelight@0: 		return;
truelight@0: 	}
truelight@0: 
ludde@2063: 	// Trim spaces.
ludde@2063: 	// After this str points to the command name, and s points to the command contents
tron@3015: 	for (t = s; t > str && (t[-1] == ' ' || t[-1] == '\t'); t--);
ludde@2063: 	*t = 0;
ludde@2063: 	s++;
truelight@0: 
peter1138@5111: 	/* Check string is valid UTF-8 */
peter1138@5111: 	{
peter1138@5111: 		const char *tmp;
peter1138@5111: 		for (tmp = s; *tmp != '\0';) {
peter1138@5111: 			size_t len = Utf8Validate(tmp);
peter1138@5111: 			if (len == 0) fatal("Invalid UTF-8 sequence in '%s'", s);
peter1138@5111: 			tmp += len;
peter1138@5111: 		}
peter1138@5111: 	}
peter1138@5111: 
ludde@2087: 	// Check if the string has a case..
ludde@2087: 	// The syntax for cases is IDENTNAME.case
ludde@2087: 	casep = strchr(str, '.');
ludde@2087: 	if (casep) *casep++ = 0;
ludde@2087: 
ludde@2063: 	// Check if this string already exists..
ludde@2063: 	ent = HashFind(str);
truelight@0: 
truelight@0: 	if (master) {
tron@3015: 		if (ent != NULL && casep == NULL) {
Darkvater@4922: 			error("String name '%s' is used multiple times", str);
truelight@0: 			return;
truelight@0: 		}
truelight@0: 
tron@3015: 		if (ent == NULL && casep != NULL) {
Darkvater@4922: 			error("Base string name '%s' doesn't exist yet. Define it before defining a case.", str);
truelight@0: 			return;
truelight@0: 		}
truelight@0: 
ludde@2087: 		if (ent == NULL) {
ludde@2087: 			if (_strings[_next_string_id]) {
Darkvater@4922: 				error("String ID 0x%X for '%s' already in use by '%s'", ent, str, _strings[_next_string_id]->name);
ludde@2087: 				return;
ludde@2087: 			}
ludde@2087: 
ludde@2087: 			// Allocate a new LangString
KUDr@5860: 			ent = CallocT<LangString>(1);
ludde@2087: 			_strings[_next_string_id] = ent;
ludde@2087: 			ent->index = _next_string_id++;
ludde@2087: 			ent->name = strdup(str);
Darkvater@2566: 			ent->line = _cur_line;
ludde@2087: 
ludde@2087: 			HashAdd(str, ent);
ludde@2087: 		}
ludde@2087: 
tron@3015: 		if (casep != NULL) {
KUDr@5860: 			Case* c = MallocT<Case>(1);
tron@3015: 
ludde@2087: 			c->caseidx = ResolveCaseName(casep, strlen(casep));
ludde@2087: 			c->string = strdup(s);
ludde@2087: 			c->next = ent->english_case;
ludde@2087: 			ent->english_case = c;
ludde@2087: 		} else {
ludde@2087: 			ent->english = strdup(s);
ludde@2087: 		}
ludde@2087: 
truelight@0: 	} else {
ludde@2087: 		if (ent == NULL) {
Darkvater@4922: 			warning("String name '%s' does not exist in master file", str);
truelight@0: 			return;
truelight@0: 		}
truelight@0: 
tron@3015: 		if (ent->translated && casep == NULL) {
Darkvater@4922: 			error("String name '%s' is used multiple times", str);
truelight@0: 			return;
truelight@0: 		}
truelight@0: 
ludde@2087: 		if (s[0] == ':' && s[1] == '\0' && casep == NULL) {
ludde@2063: 			// Special syntax :: means we should just inherit the master string
ludde@2087: 			ent->translated = strdup(ent->english);
truelight@0: 		} else {
ludde@2087: 			// make sure that the commands match
tron@3015: 			if (!CheckCommandsMatch(s, ent->english, str)) return;
ludde@2087: 
tron@3015: 			if (casep != NULL) {
KUDr@5860: 				Case* c = MallocT<Case>(1);
tron@3015: 
ludde@2087: 				c->caseidx = ResolveCaseName(casep, strlen(casep));
ludde@2087: 				c->string = strdup(s);
ludde@2087: 				c->next = ent->translated_case;
ludde@2087: 				ent->translated_case = c;
ludde@2087: 			} else {
ludde@2087: 				ent->translated = strdup(s);
ludde@2063: 			}
truelight@0: 		}
truelight@0: 	}
truelight@0: }
truelight@0: 
ludde@2063: 
ludde@2063: static void rstrip(char *buf)
truelight@0: {
ludde@2063: 	int i = strlen(buf);
tron@3015: 	while (i > 0 && (buf[i - 1] == '\r' || buf[i - 1] == '\n' || buf[i - 1] == ' ')) i--;
tron@3015: 	buf[i] = '\0';
truelight@0: }
truelight@0: 
ludde@2063: 
ludde@2063: static void ParseFile(const char *file, bool english)
tron@2059: {
ludde@2063: 	FILE *in;
truelight@0: 	char buf[2048];
truelight@0: 
tron@2459: 	_file = file;
tron@2459: 
Darkvater@5976: 	/* For each new file we parse, reset the genders, and language codes */
ludde@2087: 	_numgenders = 0;
Darkvater@5976: 	_lang_name[0] = _lang_ownname[0] = _lang_isocode[0] = '\0';
ludde@2087: 	// TODO:!! We can't reset the cases. In case the translated strings
ludde@2087: 	// derive some strings from english....
ludde@2087: 
truelight@0: 	in = fopen(file, "r");
Darkvater@4922: 	if (in == NULL) fatal("Cannot open file");
truelight@0: 	_cur_line = 1;
KUDr@6308: 	while (fgets(buf, sizeof(buf), in) != NULL) {
ludde@2063: 		rstrip(buf);
ludde@2063: 		HandleString(buf, english);
truelight@0: 		_cur_line++;
truelight@0: 	}
truelight@0: 	fclose(in);
Darkvater@5976: 
Darkvater@5976: 	if (StrEmpty(_lang_name) || StrEmpty(_lang_ownname) || StrEmpty(_lang_isocode)) {
Darkvater@5976: 		fatal("Language must include ##name, ##ownname and ##isocode");
Darkvater@5976: 	}
ludde@2063: }
tron@2059: 
tron@2059: 
ludde@2063: static uint32 MyHashStr(uint32 hash, const char *s)
ludde@2063: {
tron@3015: 	for (; *s != '\0'; s++) {
tron@3015: 		hash = ROL(hash, 3) ^ *s;
tron@4077: 		hash = (hash & 1 ? hash >> 1 ^ 0xDEADBEEF : hash >> 1);
ludde@2063: 	}
ludde@2063: 	return hash;
ludde@2063: }
truelight@0: 
ludde@2063: 
ludde@2063: // make a hash of the file to get a unique "version number"
bjarni@6298: static void MakeHashOfStrings()
ludde@2063: {
ludde@2063: 	uint32 hash = 0;
tron@3015: 	uint i;
ludde@2063: 
tron@3015: 	for (i = 0; i != lengthof(_strings); i++) {
tron@3015: 		const LangString* ls = _strings[i];
tron@3015: 
tron@3015: 		if (ls != NULL) {
tron@3015: 			const CmdStruct* cs;
tron@3015: 			const char* s;
tron@3015: 			char buf[256];
tron@3015: 			int argno;
tron@3015: 			int casei;
tron@3015: 
ludde@2087: 			s = ls->name;
ludde@2063: 			hash ^= i * 0x717239;
tron@4077: 			hash = (hash & 1 ? hash >> 1 ^ 0xDEADBEEF : hash >> 1);
ludde@2063: 			hash = MyHashStr(hash, s + 1);
ludde@2063: 
ludde@2087: 			s = ls->english;
tron@3015: 			while ((cs = ParseCommandString(&s, buf, &argno, &casei)) != NULL) {
tron@3015: 				if (cs->flags & C_DONTCOUNT) continue;
ludde@2087: 
ludde@2063: 				hash ^= (cs - _cmd_structs) * 0x1234567;
tron@4077: 				hash = (hash & 1 ? hash >> 1 ^ 0xF00BAA4 : hash >> 1);
darkvater@222: 			}
truelight@0: 		}
truelight@0: 	}
ludde@2063: 	_hash = hash;
truelight@0: }
truelight@0: 
ludde@2063: 
tron@3015: static uint CountInUse(uint grp)
tron@2059: {
truelight@0: 	int i;
truelight@0: 
tron@3015: 	for (i = 0x800; --i >= 0;) if (_strings[(grp << 11) + i] != NULL) break;
truelight@0: 	return i + 1;
truelight@0: }
truelight@0: 
truelight@0: 
ludde@2063: bool CompareFiles(const char *n1, const char *n2)
truelight@0: {
truelight@0: 	FILE *f1, *f2;
truelight@0: 	char b1[4096];
truelight@0: 	char b2[4096];
truelight@0: 	size_t l1, l2;
truelight@0: 
truelight@0: 	f2 = fopen(n2, "rb");
truelight@0: 	if (f2 == NULL) return false;
truelight@0: 
truelight@0: 	f1 = fopen(n1, "rb");
Darkvater@4922: 	if (f1 == NULL) fatal("can't open %s", n1);
truelight@0: 
truelight@0: 	do {
truelight@0: 		l1 = fread(b1, 1, sizeof(b1), f1);
truelight@0: 		l2 = fread(b2, 1, sizeof(b2), f2);
truelight@0: 
ludde@2063: 		if (l1 != l2 || memcmp(b1, b2, l1)) {
truelight@0: 			fclose(f2);
truelight@0: 			fclose(f1);
truelight@0: 			return false;
truelight@0: 		}
truelight@0: 	} while (l1);
truelight@0: 
truelight@0: 	fclose(f2);
truelight@0: 	fclose(f1);
truelight@0: 	return true;
truelight@0: }
truelight@0: 
ludde@2063: 
ludde@2063: static void WriteStringsH(const char *filename)
truelight@0: {
truelight@0: 	FILE *out;
truelight@0: 	int i;
truelight@0: 	int next = -1;
truelight@0: 	int lastgrp;
truelight@0: 
truelight@0: 	out = fopen("tmp.xxx", "w");
Darkvater@4922: 	if (out == NULL) fatal("can't open tmp.xxx");
truelight@0: 
rubidium@5838: 	fprintf(out, "enum StringIdEnum {");
truelight@0: 
truelight@0: 	lastgrp = 0;
truelight@0: 
tron@3015: 	for (i = 0; i != lengthof(_strings); i++) {
tron@3015: 		if (_strings[i] != NULL) {
tron@3015: 			fprintf(out, next == i ? "\t%s,\n" : "\n\t%s = 0x%X,\n", _strings[i]->name, i);
truelight@0: 			next = i + 1;
truelight@0: 		}
truelight@0: 	}
truelight@0: 
truelight@0: 	fprintf(out, "};\n");
truelight@0: 
darkvater@222: 	fprintf(out,
truelight@0: 		"\nenum {\n"
darkvater@181: 		"\tLANGUAGE_PACK_IDENT = 0x474E414C, // Big Endian value for 'LANG' (LE is 0x 4C 41 4E 47)\n"
truelight@0: 		"\tLANGUAGE_PACK_VERSION = 0x%X,\n"
tron@3015: 		"};\n", (uint)_hash
tron@3015: 	);
ludde@2063: 
truelight@0: 	fclose(out);
truelight@0: 
ludde@2063: 	if (CompareFiles("tmp.xxx", filename)) {
truelight@0: 		// files are equal. tmp.xxx is not needed
truelight@0: 		unlink("tmp.xxx");
truelight@0: 	} else {
truelight@0: 		// else rename tmp.xxx into filename
Darkvater@2482: #if defined(WIN32) || defined(WIN64)
truelight@0: 		unlink(filename);
truelight@0: #endif
Darkvater@4922: 		if (rename("tmp.xxx", filename) == -1) fatal("rename() failed");
truelight@0: 	}
truelight@0: }
truelight@0: 
ludde@2087: static int TranslateArgumentIdx(int argidx)
ludde@2063: {
ludde@2063: 	int i, sum;
ludde@2063: 
ludde@2082: 	if (argidx < 0 || argidx >= lengthof(_cur_pcs.cmd))
Darkvater@4922: 		fatal("invalid argidx %d", argidx);
ludde@2063: 
tron@2952: 	for (i = sum = 0; i < argidx; i++) {
tron@2374: 		const CmdStruct *cs = _cur_pcs.cmd[i];
tron@3015: 		sum += (cs != NULL) ? cs->consumes : 1;
ludde@2063: 	}
ludde@2063: 
ludde@2082: 	return sum;
ludde@2082: }
ludde@2082: 
bjarni@6298: static void PutArgidxCommand()
ludde@2082: {
peter1138@5108: 	PutUtf8(SCC_ARG_INDEX);
ludde@2087: 	PutByte(TranslateArgumentIdx(_cur_argidx));
ludde@2087: }
ludde@2087: 
ludde@2087: 
ludde@2087: static void PutCommandString(const char *str)
ludde@2087: {
ludde@2087: 	const CmdStruct *cs;
ludde@2087: 	char param[256];
ludde@2087: 	int argno;
ludde@2087: 	int casei;
ludde@2087: 
ludde@2087: 	_cur_argidx = 0;
ludde@2087: 
ludde@2087: 	while (*str != '\0') {
ludde@2087: 		// Process characters as they are until we encounter a {
ludde@2087: 		if (*str != '{') {
ludde@2087: 			PutByte(*str++);
ludde@2087: 			continue;
ludde@2087: 		}
ludde@2087: 		cs = ParseCommandString(&str, param, &argno, &casei);
ludde@2087: 		if (cs == NULL) break;
ludde@2087: 
ludde@2087: 		if (casei != -1) {
peter1138@5108: 			PutUtf8(SCC_SETCASE); // {SETCASE}
ludde@2087: 			PutByte(casei);
ludde@2087: 		}
ludde@2087: 
ludde@2087: 		// For params that consume values, we need to handle the argindex properly
tron@3015: 		if (cs->consumes > 0) {
ludde@2087: 			// Check if we need to output a move-param command
tron@3015: 			if (argno != -1 && argno != _cur_argidx) {
ludde@2087: 				_cur_argidx = argno;
ludde@2087: 				PutArgidxCommand();
ludde@2087: 			}
ludde@2087: 
ludde@2087: 			// Output the one from the master string... it's always accurate.
ludde@2087: 			cs = _cur_pcs.cmd[_cur_argidx++];
tron@3015: 			if (cs == NULL) {
Darkvater@4922: 				fatal("%s: No argument exists at position %d", _cur_ident, _cur_argidx - 1);
tron@3015: 			}
ludde@2087: 		}
ludde@2087: 
ludde@2087: 		cs->proc(param, cs->value);
ludde@2087: 	}
ludde@2087: }
ludde@2087: 
ludde@2087: static void WriteLength(FILE *f, uint length)
ludde@2087: {
ludde@2087: 	if (length < 0xC0) {
ludde@2087: 		fputc(length, f);
ludde@2087: 	} else if (length < 0x4000) {
ludde@2087: 		fputc((length >> 8) | 0xC0, f);
ludde@2087: 		fputc(length & 0xFF, f);
ludde@2087: 	} else {
Darkvater@4922: 		fatal("string too long");
ludde@2087: 	}
ludde@2063: }
ludde@2063: 
ludde@2063: 
ludde@2063: static void WriteLangfile(const char *filename, int show_todo)
truelight@0: {
truelight@0: 	FILE *f;
tron@3015: 	uint in_use[32];
truelight@0: 	LanguagePackHeader hdr;
tron@3015: 	uint i;
tron@3015: 	uint j;
truelight@0: 
truelight@0: 	f = fopen(filename, "wb");
Darkvater@4922: 	if (f == NULL) fatal("can't open %s", filename);
truelight@0: 
truelight@0: 	memset(&hdr, 0, sizeof(hdr));
tron@2952: 	for (i = 0; i != 32; i++) {
tron@3015: 		uint n = CountInUse(i);
tron@3015: 
truelight@0: 		in_use[i] = n;
truelight@0: 		hdr.offsets[i] = TO_LE16(n);
truelight@0: 	}
truelight@0: 
darkvater@222: 	// see line 655: fprintf(..."\tLANGUAGE_PACK_IDENT = 0x474E414C,...)
darkvater@181: 	hdr.ident = TO_LE32(0x474E414C); // Big Endian value for 'LANG'
truelight@0: 	hdr.version = TO_LE32(_hash);
ludde@2082: 	hdr.plural_form = _lang_pluralform;
truelight@0: 	strcpy(hdr.name, _lang_name);
truelight@0: 	strcpy(hdr.own_name, _lang_ownname);
miham@1376: 	strcpy(hdr.isocode, _lang_isocode);
truelight@0: 
truelight@0: 	fwrite(&hdr, sizeof(hdr), 1, f);
truelight@0: 
tron@2952: 	for (i = 0; i != 32; i++) {
tron@2952: 		for (j = 0; j != in_use[i]; j++) {
tron@3015: 			const LangString* ls = _strings[(i << 11) + j];
tron@3015: 			const Case* casep;
tron@3015: 			const char* cmdp;
truelight@0: 
ludde@2063: 			// For undefined strings, just set that it's an empty string
ludde@2087: 			if (ls == NULL) {
ludde@2063: 				WriteLength(f, 0);
ludde@2063: 				continue;
ludde@2063: 			}
truelight@0: 
ludde@2087: 			_cur_ident = ls->name;
Darkvater@2566: 			_cur_line = ls->line;
ludde@2082: 
ludde@2063: 			// Produce a message if a string doesn't have a translation.
tron@3015: 			if (show_todo > 0 && ls->translated == NULL) {
ludde@2063: 				if (show_todo == 2) {
Darkvater@4922: 					warning("'%s' is untranslated", ls->name);
ludde@2063: 				} else {
ludde@2063: 					const char *s = "<TODO> ";
tron@3015: 					while (*s != '\0') PutByte(*s++);
ludde@2063: 				}
ludde@2063: 			}
ludde@2063: 
ludde@2063: 			// Extract the strings and stuff from the english command string
ludde@2087: 			ExtractCommandString(&_cur_pcs, ls->english, false);
ludde@2063: 
tron@3015: 			if (ls->translated_case != NULL || ls->translated != NULL) {
ludde@2087: 				casep = ls->translated_case;
ludde@2087: 				cmdp = ls->translated;
ludde@2087: 			} else {
ludde@2087: 				casep = ls->english_case;
ludde@2087: 				cmdp = ls->english;
ludde@2087: 			}
ludde@2063: 
Darkvater@2586: 			_translated = _masterlang || (cmdp != ls->english);
Darkvater@2585: 
tron@3015: 			if (casep != NULL) {
tron@3015: 				const Case* c;
tron@3015: 				uint num;
tron@3015: 
ludde@2087: 				// Need to output a case-switch.
ludde@2087: 				// It has this format
ludde@2087: 				// <0x9E> <NUM CASES> <CASE1> <LEN1> <STRING1> <CASE2> <LEN2> <STRING2> <CASE3> <LEN3> <STRING3> <STRINGDEFAULT>
ludde@2087: 				// Each LEN is printed using 2 bytes in big endian order.
peter1138@5108: 				PutUtf8(SCC_SWITCH_CASE);
ludde@2087: 				// Count the number of cases
tron@2952: 				for (num = 0, c = casep; c; c = c->next) num++;
ludde@2087: 				PutByte(num);
ludde@2087: 
ludde@2087: 				// Write each case
tron@3015: 				for (c = casep; c != NULL; c = c->next) {
ludde@2087: 					int pos;
tron@3015: 
ludde@2087: 					PutByte(c->caseidx);
ludde@2087: 					// Make some space for the 16-bit length
ludde@2087: 					pos = _put_pos;
ludde@2087: 					PutByte(0);
ludde@2087: 					PutByte(0);
ludde@2087: 					// Write string
ludde@2087: 					PutCommandString(c->string);
ludde@2087: 					PutByte(0); // terminate with a zero
ludde@2087: 					// Fill in the length
tron@3015: 					_put_buf[pos + 0] = GB(_put_pos - (pos + 2), 8, 8);
tron@3015: 					_put_buf[pos + 1] = GB(_put_pos - (pos + 2), 0, 8);
truelight@0: 				}
ludde@2087: 			}
truelight@0: 
tron@3015: 			if (cmdp != NULL) PutCommandString(cmdp);
truelight@0: 
ludde@2063: 			WriteLength(f, _put_pos);
ludde@2063: 			fwrite(_put_buf, 1, _put_pos, f);
ludde@2063: 			_put_pos = 0;
truelight@0: 		}
truelight@0: 	}
truelight@0: 
truelight@0: 	fputc(0, f);
truelight@0: 	fclose(f);
truelight@0: }
truelight@0: 
truelight@4370: /** Multi-OS mkdirectory function */
truelight@4370: static inline void ottd_mkdir(const char *directory)
truelight@4370: {
truelight@4370: #if defined(WIN32) || defined(__WATCOMC__)
truelight@4370: 		mkdir(directory);
truelight@4370: #else
truelight@4370: 		mkdir(directory, 0755);
truelight@4370: #endif
truelight@4370: }
truelight@4370: 
truelight@4370: /** Create a path consisting of an already existing path, a possible
truelight@4370:  * path seperator and the filename. The seperator is only appended if the path
truelight@4370:  * does not already end with a seperator */
truelight@4370: static inline char *mkpath(char *buf, size_t buflen, const char *path, const char *file)
truelight@4370: {
truelight@4370: 	char *p;
truelight@4370: 	ttd_strlcpy(buf, path, buflen); // copy directory into buffer
truelight@4370: 
truelight@4370: 	p = strchr(buf, '\0'); // add path seperator if necessary
truelight@4370: 	if (p[-1] != PATHSEPCHAR && (size_t)(p - buf) + 1 < buflen) *p++ = PATHSEPCHAR;
truelight@4370: 	ttd_strlcpy(p, file, buflen - (size_t)(p - buf)); // catenate filename at end of buffer
truelight@4370: 	return buf;
truelight@4370: }
truelight@4370: 
rubidium@5754: #if defined(__MINGW32__)
truelight@4379: /**
truelight@4379:  * On MingW, it is common that both / as \ are accepted in the
truelight@4379:  * params. To go with those flow, we rewrite all incoming /
truelight@4379:  * simply to \, so internally we can safely assume \.
truelight@4379:  */
truelight@4379: static inline char *replace_pathsep(char *s)
truelight@4379: {
truelight@4379: 	char *c;
truelight@4379: 
truelight@4379: 	for (c = s; *c != '\0'; c++) if (*c == '/') *c = '\\';
truelight@4379: 	return s;
truelight@4379: }
truelight@4379: #else
truelight@4379: static inline char *replace_pathsep(char *s) { return s; }
truelight@4379: #endif
ludde@2063: 
darkvater@250: int CDECL main(int argc, char* argv[])
truelight@0: {
truelight@4370: 	char pathbuf[256];
glx@4461: 	const char *src_dir = ".";
glx@4461: 	const char *dest_dir = NULL;
truelight@4370: 
truelight@0: 	int show_todo = 0;
truelight@0: 
glx@4461: 	while (argc > 1 && *argv[1] == '-') {
glx@4461: 		if (strcmp(argv[1], "-v") == 0 || strcmp(argv[1], "--version") == 0) {
glx@4461: 			puts("$Revision$");
glx@4461: 			return 0;
glx@4461: 		}
truelight@0: 
glx@4461: 		if (strcmp(argv[1], "-t") == 0 || strcmp(argv[1], "--todo") == 0) {
glx@4461: 			show_todo = 1;
glx@4461: 			argc--, argv++;
glx@4461: 			continue;
glx@4461: 		}
truelight@0: 
glx@4461: 		if (strcmp(argv[1], "-w") == 0 || strcmp(argv[1], "--warning") == 0) {
glx@4461: 			show_todo = 2;
glx@4461: 			argc--, argv++;
glx@4461: 			continue;
glx@4461: 		}
glx@4461: 
glx@4461: 		if (strcmp(argv[1], "-h") == 0 || strcmp(argv[1], "--help") == 0 || strcmp(argv[1], "-?") == 0) {
glx@4461: 			puts(
glx@4461: 				"strgen - $Revision$\n"
glx@4461: 				" -v | --version    print version information and exit\n"
glx@4461: 				" -t | --todo       replace any untranslated strings with '<TODO>'\n"
glx@4461: 				" -w | --warning    print a warning for any untranslated strings\n"
glx@4461: 				" -h | -? | --help  print this help message and exit\n"
glx@4461: 				" -s | --source_dir search for english.txt in the specified directory\n"
glx@4461: 				" -d | --dest_dir   put output file in the specified directory, create if needed\n"
glx@4461: 				" Run without parameters and strgen will search for english.txt and parse it,\n"
glx@4461: 				" creating strings.h. Passing an argument, strgen will translate that language\n"
glx@4461: 				" file using english.txt as a reference and output <language>.lng."
glx@4461: 			);
glx@4461: 			return 0;
glx@4461: 		}
glx@4461: 
glx@4461: 		if (argc > 2 && (strcmp(argv[1], "-s") == 0 || strcmp(argv[1], "--source_dir") == 0)) {
glx@4461: 			src_dir = replace_pathsep(argv[2]);
glx@4461: 			argc -= 2, argv += 2;
glx@4464: 			continue;
glx@4461: 		}
glx@4461: 
glx@4461: 		if (argc > 2 && (strcmp(argv[1], "-d") == 0 || strcmp(argv[1], "--dest_dir") == 0)) {
glx@4461: 			dest_dir = replace_pathsep(argv[2]);
glx@4461: 			argc -= 2, argv += 2;
glx@4464: 			continue;
glx@4461: 		}
glx@4464: 
glx@4464: 		fprintf(stderr, "Invalid arguments\n");
glx@4464: 		return 0;
truelight@0: 	}
truelight@0: 
glx@4461: 	if (dest_dir == NULL) dest_dir = src_dir; // if dest_dir is not specified, it equals src_dir
truelight@4370: 
truelight@4370: 	/* strgen has two modes of operation. If no (free) arguments are passed
truelight@4370: 	 * strgen generates strings.h to the destination directory. If it is supplied
truelight@4370: 	 * with a (free) parameter the program will translate that language to destination
truelight@4370: 	 * directory. As input english.txt is parsed from the source directory */
truelight@0: 	if (argc == 1) {
truelight@4370: 		mkpath(pathbuf, lengthof(pathbuf), src_dir, "english.txt");
truelight@4370: 
truelight@4370: 		/* parse master file */
Darkvater@2586: 		_masterlang = true;
truelight@4370: 		ParseFile(pathbuf, true);
ludde@2063: 		MakeHashOfStrings();
ludde@2063: 		if (_errors) return 1;
truelight@0: 
truelight@4370: 		/* write strings.h */
truelight@4370: 		ottd_mkdir(dest_dir);
truelight@4370: 		mkpath(pathbuf, lengthof(pathbuf), dest_dir, "strings.h");
truelight@4370: 		WriteStringsH(pathbuf);
tron@3015: 	} else if (argc == 2) {
truelight@4370: 		char *r;
ludde@2063: 
truelight@4370: 		mkpath(pathbuf, lengthof(pathbuf), src_dir, "english.txt");
truelight@4370: 
truelight@4370: 		/* parse master file and check if target file is correct */
Darkvater@2586: 		_masterlang = false;
truelight@4370: 		ParseFile(pathbuf, true);
ludde@2063: 		MakeHashOfStrings();
truelight@4379: 		ParseFile(replace_pathsep(argv[1]), false); // target file
ludde@2063: 		if (_errors) return 1;
ludde@2063: 
truelight@4370: 		/* get the targetfile, strip any directories and append to destination path */
truelight@4370: 		r = strrchr(argv[1], PATHSEPCHAR);
truelight@4370: 		mkpath(pathbuf, lengthof(pathbuf), dest_dir, (r != NULL) ? &r[1] : argv[1]);
truelight@4370: 
truelight@4370: 		/* rename the .txt (input-extension) to .lng */
truelight@4370: 		r = strrchr(pathbuf, '.');
truelight@4370: 		if (r == NULL || strcmp(r, ".txt") != 0) r = strchr(pathbuf, '\0');
truelight@4370: 		ttd_strlcpy(r, ".lng", (size_t)(r - pathbuf));
truelight@4370: 		WriteLangfile(pathbuf, show_todo);
Darkvater@5548: 
Darkvater@5548: 		/* if showing warnings, print a summary of the language */
Darkvater@5548: 		if (show_todo == 2) {
Darkvater@5693: 			fprintf(stdout, "%d warnings and %d errors for %s\n", _warnings, _errors, pathbuf);
Darkvater@5548: 		}
truelight@0: 	} else {
truelight@4370: 		fprintf(stderr, "Invalid arguments\n");
truelight@0: 	}
truelight@0: 
truelight@0: 	return 0;
truelight@0: }