Mercurial > hg > openttd

changeset 2082:ca9bda54f880 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
(svn r2592) Feature: [strgen] New way to specify plural forms. - {NUM} {PLURAL 0 car cars}: Prints either car or cars depending on if the argument 0 is plural. Also supports languages with weird plural forms such as Polish. The plural format needs to be specified in the beginning of the langfile, like "##plural 7" for Polish.
author ludde <ludde@openttd.org>
date Sat, 16 Jul 2005 17:12:32 +0000
parents 960e61968bab
children 864c2a190fbc
files strgen/strgen.c strings.c
diffstat 2 files changed, 230 insertions(+), 20 deletions(-) [+]
line wrap: on
line diff
--- a/strgen/strgen.c
+++ b/strgen/strgen.c
@@ -28,6 +28,8 @@
 	char own_name[32];	// the localized name of this language
 	char isocode[16];	// the ISO code for the language (not country code)
 	uint16 offsets[32];	// the offsets
+	byte plural_form;		// plural form index
+	byte pad[3];				// pad header to be a multiple of 4
 } LanguagePackHeader;
 
 typedef struct CmdStruct {
@@ -35,6 +37,7 @@
 	ParseCmdProc proc;
 	long value;
 	int8 consumes;
+	bool dont_count;
 } CmdStruct;
 
 static int _cur_line;
@@ -54,7 +57,12 @@
 
 static uint32 _hash;
 static char _lang_name[32], _lang_ownname[32], _lang_isocode[16];
+static byte _lang_pluralform;
 
+// for each plural value, this is the number of plural forms.
+static const byte _plural_form_counts[] = { 2,1,2,3,3,3,3,3,4 };
+
+static const char *_cur_ident;
 
 static uint HashStr(const char *s)
 {
@@ -169,7 +177,7 @@
 	int x,y;
 
 	x = strtol(buf, &err, 0);
-	if (*err != 0) Fatal("SetXY param invalid");
+	if (*err != ' ') Fatal("SetXY param invalid");
 	y = strtol(err+1, &err, 0);
 	if (*err != 0) Fatal("SetXY param invalid");
 
@@ -178,6 +186,102 @@
 	PutByte((byte)y);
 }
 
+// The plural specifier looks like
+// {NUM} {PLURAL -1 passenger passengers} then it picks either passenger/passengers depending on the count in NUM
+
+// This is encoded like
+//  CommandByte <ARG#> <NUM> {Length of each string} {each string}
+
+bool ParseRelNum(char **buf, int *value, bool *relative)
+{
+	char *s = *buf, *end;
+	bool rel = false;
+	while (*s == ' ' || *s == '\t') s++;
+	if (*s == '+') { rel = true; s++; }
+	*value = strtol(s, &end, 0);
+	if (end == s) return false;
+	*relative = rel | (*value < 0);
+	*buf = end;
+	return true;
+}
+
+// Parse out the next word, or NULL
+char *ParseWord(char **buf)
+{
+	char *s = *buf, *r;
+	while (*s == ' ' || *s == '\t') s++;
+	if (*s == 0)
+		return NULL;
+
+	if (*s == '"') {
+		r = ++s;
+		// parse until next " or NUL
+		for(;;) {
+			if (*s == 0)
+				break;
+			if (*s == '"') {
+				*s++ = 0;
+				break;
+			}
+			s++;
+		}
+	} else {
+		// proceed until whitespace or NUL
+		r = s;
+		for(;;) {
+			if (*s == 0)
+				break;
+			if (*s == ' ' || *s == '\t') {
+				*s++ = 0;
+				break;
+			}
+			s++;
+		}
+	}
+	*buf = s;
+	return r;
+}
+
+// Forward declaration
+static int TranslateArgumentIdx(int arg, bool relative);
+
+static void EmitPlural(char *buf, int value)
+{
+	int v,i,j;
+	bool relative;
+	char *words[5];
+	int nw = 0;
+
+	// Parse out the number.
+	if (!ParseRelNum(&buf, &v, &relative))
+		Fatal("Plural param invalid");
+
+	// Parse each string
+	for(nw=0; nw<5; nw++) {
+		words[nw] = ParseWord(&buf);
+		if (!words[nw])
+			break;
+	}
+
+	if (nw == 0)
+		Fatal("No plural words");
+
+	if (_plural_form_counts[_lang_pluralform] != nw)
+		Fatal("%s: Invalid number of plural forms. Expecting %d, found %d.", _cur_ident,
+			_plural_form_counts[_lang_pluralform], nw);
+
+	PutByte(0x7D);
+	PutByte(TranslateArgumentIdx(v, relative));
+	PutByte(nw);
+	for(i=0; i<nw; i++)
+		PutByte(strlen(words[i]));
+	for(i=0; i<nw; i++) {
+		for(j=0; words[i][j]; j++)
+			PutByte(words[i][j]);
+	}
+}
+
+
 
 static const CmdStruct _cmd_structs[] = {
 	// Update position
@@ -189,7 +293,7 @@
 	{"BIGFONT", EmitSingleByte, 9, 0},
 
 	// New line
-	{"", EmitSingleByte, 10, 0},
+	{"", EmitSingleByte, 10, 0, true},
 
 	// Colors
 	{"BLUE", EmitSingleByte,    15, 0},
@@ -243,6 +347,8 @@
 	{"STATIONFEATURES", EmitEscapedByte, 10, 1},				// station features string, icons of the features
 	{"INDUSTRY", EmitEscapedByte, 11, 1},			// industry, takes an industry #
 
+	{"PLURAL", EmitPlural, 0, 0, true},							// plural specifier
+
 	{"DATE_LONG", EmitSingleByte, 0x82, 1},
 	{"DATE_SHORT", EmitSingleByte, 0x83, 1},
 
@@ -349,7 +455,7 @@
 				Error("Missing } from command '%s'", start);
 				return NULL;
 			}
-			if ( s - start == 60)
+			if ( s - start == 250)
 				Fatal("param command too long");
 			*param++ = c;
 		}
@@ -372,6 +478,10 @@
 		ttd_strlcpy(_lang_ownname, str + 8, sizeof(_lang_ownname));
 	} else if (!memcmp(str, "isocode ", 8)) {
 		ttd_strlcpy(_lang_isocode, str + 8, sizeof(_lang_isocode));
+	} else if (!memcmp(str, "plural ", 7)) {
+		_lang_pluralform = atoi(str + 7);
+		if (_lang_pluralform >= lengthof(_plural_form_counts))
+			Fatal("Invalid pluralform %d", _lang_pluralform);
 	} else {
 		Fatal("unknown pragma '%s'", str);
 	}
@@ -414,7 +524,7 @@
 			if (p->cmd[argidx] != NULL && p->cmd[argidx] != ar) Fatal("duplicate param idx %d", argidx);
 
 			p->cmd[argidx++] = ar;
-		} else if (ar->cmd[0] != '\0') { // Ignore {}.. it can appear in any order.
+		} else if (!ar->dont_count) { // Ignore some of them
 			if (p->np >= lengthof(p->pairs)) Fatal("too many commands in string, max %d", lengthof(p->pairs));
 			p->pairs[p->np].a = ar;
 			p->pairs[p->np].v = param[0]?strdup(param):"";
@@ -604,7 +714,7 @@
 	uint32 hash = 0;
 	char *s;
 	const CmdStruct *cs;
-	char buf[128];
+	char buf[256];
 	int i;
 	int argno;
 
@@ -729,21 +839,31 @@
 	}
 }
 
+static ParsedCommandStruct _cur_pcs;
+static int _cur_argidx;
 
-static void PutArgidxCommand(ParsedCommandStruct *pcs, int argidx)
+static int TranslateArgumentIdx(int argidx, bool relative)
 {
 	int i, sum;
 
-	if (argidx >= lengthof(pcs->cmd))
+	if (relative)
+		argidx += _cur_argidx;
+
+	if (argidx < 0 || argidx >= lengthof(_cur_pcs.cmd))
 		Fatal("invalid argidx %d", argidx);
 
 	for(i = sum = 0; i < argidx; i++) {
-		const CmdStruct *cs = pcs->cmd[i++];
+		const CmdStruct *cs = _cur_pcs.cmd[i++];
 		sum += cs ? cs->consumes : 1;
 	}
 
+	return sum;
+}
+
+static void PutArgidxCommand(void)
+{
 	PutByte(0x7C);
-	PutByte((byte)sum);
+	PutByte(TranslateArgumentIdx(0, true));
 }
 
 
@@ -754,9 +874,8 @@
 	LanguagePackHeader hdr;
 	int i,j;
 	const CmdStruct *cs;
-	char param[128];
+	char param[256];
 	int argno;
-	ParsedCommandStruct pcs;
 
 	f = fopen(filename, "wb");
 	if (f == NULL) Fatal("can't open %s", filename);
@@ -771,6 +890,7 @@
 	// see line 655: fprintf(..."\tLANGUAGE_PACK_IDENT = 0x474E414C,...)
 	hdr.ident = TO_LE32(0x474E414C); // Big Endian value for 'LANG'
 	hdr.version = TO_LE32(_hash);
+	hdr.plural_form = _lang_pluralform;
 	strcpy(hdr.name, _lang_name);
 	strcpy(hdr.own_name, _lang_ownname);
 	strcpy(hdr.isocode, _lang_isocode);
@@ -781,7 +901,6 @@
 		for(j = 0; j != in_use[i]; j++) {
 			int idx = (i<<11)+j;
 			char *str;
-			int argidx;
 
 			// For undefined strings, just set that it's an empty string
 			if (_strname[idx] == NULL) {
@@ -789,6 +908,8 @@
 				continue;
 			}
 
+			_cur_ident = _strname[idx];
+
 			// Produce a message if a string doesn't have a translation.
 			if (show_todo && _translated[idx] == NULL) {
 				if (show_todo == 2) {
@@ -800,10 +921,10 @@
 			}
 
 			// Extract the strings and stuff from the english command string
-			ExtractCommandString(&pcs, _master[idx], false);
+			ExtractCommandString(&_cur_pcs, _master[idx], false);
 
 			str = _translated[idx] ? _translated[idx] : _master[idx];
-			argidx = 0;
+			_cur_argidx = 0;
 
 			while (*str != '\0') {
 				// Process characters as they are until we encounter a {
@@ -817,13 +938,13 @@
 				// For params that consume values, we need to handle the argindex properly
 				if (cs->consumes) {
 					// Check if we need to output a move-param command
-					if (argno!=-1 && argno != argidx) {
-						argidx = argno;
-						PutArgidxCommand(&pcs, argidx);
+					if (argno!=-1 && argno != _cur_argidx) {
+						_cur_argidx = argno;
+						PutArgidxCommand();
 					}
 
 					// Output the one from the master string... it's always accurate.
-					cs = pcs.cmd[argidx++];
+					cs = _cur_pcs.cmd[_cur_argidx++];
 					if (!cs)
 						Fatal("cs == NULL");
 				}
--- a/strings.c
+++ b/strings.c
@@ -27,6 +27,8 @@
 	char own_name[32];	// the localized name of this language
 	char isocode[16];	// the ISO code for the language (not country code)
 	uint16 offsets[32];	// the offsets
+	byte plural_form;		// how to compute plural forms
+	byte pad[3];				// pad header to be a multiple of 4
 	char data[VARARRAY_SIZE];
 } LanguagePack;
 
@@ -418,6 +420,89 @@
 	return buff;
 }
 
+static int DeterminePluralForm(int32 n)
+{
+	// The absolute value determines plurality
+	if (n < 0) n = -n;
+
+	switch(_langpack->plural_form) {
+	// Two forms, singular used for one only
+	// Used in:
+	//   Danish, Dutch, English, German, Norwegian, Swedish, Estonian, Finnish,
+	//   Greek, Hebrew, Italian, Portuguese, Spanish, Esperanto
+	case 0:
+	default:
+		return n != 1;
+
+	// Only one form
+	// Used in:
+	//   Hungarian, Japanese, Korean, Turkish
+	case 1:
+		return 0;
+
+	// Two forms, singular used for zero and one
+	// Used in:
+	//   French, Brazilian Portuguese
+	case 2:
+		return n > 1;
+
+	// Three forms, special case for zero
+	// Used in:
+	//   Latvian
+	case 3:
+		return n%10==1 && n%100!=11 ? 0 : n != 0 ? 1 : 2;
+
+	// Three forms, special case for one and two
+	// Used in:
+	//   Gaelige (Irish)
+	case 4:
+		return n==1 ? 0 : n==2 ? 1 : 2;
+
+	// Three forms, special case for numbers ending in 1[2-9]
+	// Used in:
+	//   Lithuanian
+	case 5:
+		return n%10==1 && n%100!=11 ? 0 : n%10>=2 && (n%100<10 || n%100>=20) ? 1 : 2;
+
+	// Three forms, special cases for numbers ending in 1 and 2, 3, 4, except those ending in 1[1-4]
+	// Used in:
+	//   Croatian, Czech, Russian, Slovak, Ukrainian
+	case 6:
+		return n%10==1 && n%100!=11 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2;
+
+	// Three forms, special case for one and some numbers ending in 2, 3, or 4
+	// Used in:
+	//   Polish
+	case 7:
+		return n==1 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2;
+
+	// Four forms, special case for one and all numbers ending in 02, 03, or 04
+	// Used in:
+	//   Slovenian
+	case 8:
+		return n%100==1 ? 0 : n%100==2 ? 1 : n%100==3 || n%100==4 ? 2 : 3;
+	}
+}
+
+static const char *ParseStringChoice(const char *b, uint form, char *dst, int *dstlen)
+{
+	//<NUM> {Length of each string} {each string}
+	uint n = (byte)*b++;
+	uint pos,i, mylen=0,mypos=0;
+	for(i=pos=0; i!=n; i++) {
+		uint len = (byte)*b++;
+		if (i == form) {
+			mypos = pos;
+			mylen = len;
+		}
+		pos += len;
+	}
+	*dstlen = mylen;
+	memcpy(dst, b + mypos, mylen);
+	return b + pos;
+}
+
+
 static char *FormatString(char *buff, const char *str, const int32 *argv)
 {
 	byte b;
@@ -440,9 +525,13 @@
 		case 0x7C: // Move argument pointer
 			argv = argv_orig + (byte)*str++;
 			break;
-		case 0x7D:
-			assert(0);
+		case 0x7D: { // {PLURAL}
+			int32 v = argv_orig[(byte)*str++]; // contains the number that determines plural
+			int len;
+			str = ParseStringChoice(str, DeterminePluralForm(v), buff, &len);
+			buff += len;
 			break;
+		}
 		case 0x7E: // {NUMU16}, {INT32}
 			buff = FormatNoCommaNumber(buff, GetInt32(&argv));
 			break;