Darkvater@5165: /* $Id$ */
Darkvater@5165: 
rubidium@9111: /** @file string.cpp Handling of C-type strings (char*). */
belugas@6420: 
Darkvater@5165: #include "stdafx.h"
Darkvater@5165: #include "openttd.h"
rubidium@7257: #include "debug.h"
rubidium@8130: #include "core/alloc_func.hpp"
rubidium@10299: #include "core/math_func.hpp"
rubidium@8214: #include "string_func.h"
Darkvater@5165: 
rubidium@8264: #include "table/control_codes.h"
rubidium@8264: 
Darkvater@5165: #include <stdarg.h>
Darkvater@5165: #include <ctype.h> // required for tolower()
Darkvater@5165: 
rubidium@10300: /**
rubidium@10300:  * Safer implementation of vsnprintf; same as vsnprintf except:
rubidium@10300:  * - last instead of size, i.e. replace sizeof with lastof.
rubidium@10300:  * - return gives the amount of characters added, not what it would add.
rubidium@10300:  * @param str    buffer to write to up to last
rubidium@10300:  * @param last   last character we may write to
rubidium@10300:  * @param format the formatting (see snprintf)
rubidium@10300:  * @param ap     the list of arguments for the format
rubidium@10300:  * @return the number of added characters
rubidium@10300:  */
rubidium@10300: static int CDECL vseprintf(char *str, const char *last, const char *format, va_list ap)
rubidium@10300: {
rubidium@10300: 	if (str >= last) return 0;
rubidium@10300: 	size_t size = last - str;
rubidium@10300: 	return min((int)size, vsnprintf(str, size, format, ap));
rubidium@10300: }
rubidium@10300: 
Darkvater@5165: void ttd_strlcat(char *dst, const char *src, size_t size)
Darkvater@5165: {
Darkvater@5165: 	assert(size > 0);
skidd13@10303: 	while (size > 0 && *dst != '\0') {
skidd13@10303: 		size--;
skidd13@10303: 		dst++;
skidd13@10303: 	}
skidd13@10303: 
skidd13@10303: 	ttd_strlcpy(dst, src, size);
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: 
Darkvater@5165: void ttd_strlcpy(char *dst, const char *src, size_t size)
Darkvater@5165: {
Darkvater@5165: 	assert(size > 0);
skidd13@10303: 	while (--size > 0 && *src != '\0') {
skidd13@10303: 		*dst++ = *src++;
skidd13@10303: 	}
Darkvater@5165: 	*dst = '\0';
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: 
Darkvater@5165: char* strecat(char* dst, const char* src, const char* last)
Darkvater@5165: {
Darkvater@5165: 	assert(dst <= last);
skidd13@10303: 	while (*dst != '\0') {
Darkvater@5165: 		if (dst == last) return dst;
skidd13@10303: 		dst++;
skidd13@10303: 	}
skidd13@10303: 
Darkvater@5165: 	return strecpy(dst, src, last);
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: 
Darkvater@5165: char* strecpy(char* dst, const char* src, const char* last)
Darkvater@5165: {
Darkvater@5165: 	assert(dst <= last);
skidd13@10303: 	while (dst != last && *src != '\0') {
skidd13@10303: 		*dst++ = *src++;
skidd13@10303: 	}
Darkvater@5165: 	*dst = '\0';
skidd13@10303: 
Darkvater@5165: 	if (dst == last && *src != '\0') {
rubidium@7257: #ifdef STRGEN
Darkvater@5165: 		error("String too long for destination buffer");
rubidium@7257: #else /* STRGEN */
rubidium@7257: 		DEBUG(misc, 0, "String too long for destination buffer");
rubidium@7257: 		*dst = '\0';
rubidium@7257: #endif /* STRGEN */
Darkvater@5165: 	}
Darkvater@5165: 	return dst;
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: 
rubidium@10299: char *CDECL str_fmt(const char *str, ...)
Darkvater@5165: {
Darkvater@5165: 	char buf[4096];
Darkvater@5165: 	va_list va;
Darkvater@5165: 
Darkvater@5165: 	va_start(va, str);
rubidium@10299: 	int len = vseprintf(buf, lastof(buf), str, va);
Darkvater@5165: 	va_end(va);
rubidium@10299: 	char *p = MallocT<char>(len + 1);
rubidium@10299: 	memcpy(p, buf, len + 1);
Darkvater@5165: 	return p;
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: 
Darkvater@5165: void str_validate(char *str)
Darkvater@5165: {
Darkvater@5165: 	char *dst = str;
Darkvater@5165: 	WChar c;
Darkvater@5165: 	size_t len;
Darkvater@5165: 
Darkvater@5165: 	for (len = Utf8Decode(&c, str); c != '\0'; len = Utf8Decode(&c, str)) {
Darkvater@5165: 		if (IsPrintable(c) && (c < SCC_SPRITE_START || c > SCC_SPRITE_END ||
Darkvater@5165: 			IsValidChar(c - SCC_SPRITE_START, CS_ALPHANUMERAL))) {
Darkvater@5165: 			/* Copy the character back. Even if dst is current the same as str
Darkvater@5165: 			 * (i.e. no characters have been changed) this is quicker than
Darkvater@5165: 			 * moving the pointers ahead by len */
Darkvater@5165: 			do {
Darkvater@5165: 				*dst++ = *str++;
Darkvater@5165: 			} while (--len != 0);
Darkvater@5165: 		} else {
Darkvater@5165: 			/* Replace the undesirable character with a question mark */
Darkvater@5165: 			str += len;
Darkvater@5165: 			*dst++ = '?';
Darkvater@5165: 		}
Darkvater@5165: 	}
Darkvater@5165: 
Darkvater@5165: 	*dst = '\0';
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: 
Darkvater@5165: void str_strip_colours(char *str)
Darkvater@5165: {
Darkvater@5165: 	char *dst = str;
Darkvater@5165: 	WChar c;
Darkvater@5165: 	size_t len;
Darkvater@5165: 
Darkvater@5165: 	for (len = Utf8Decode(&c, str); c != '\0'; len = Utf8Decode(&c, str)) {
Darkvater@5165: 		if (c < SCC_BLUE || c > SCC_BLACK) {
Darkvater@5165: 			/* Copy the character back. Even if dst is current the same as str
Darkvater@5165: 			 * (i.e. no characters have been changed) this is quicker than
Darkvater@5165: 			 * moving the pointers ahead by len */
Darkvater@5165: 			do {
Darkvater@5165: 				*dst++ = *str++;
Darkvater@5165: 			} while (--len != 0);
Darkvater@5165: 		} else {
Darkvater@5165: 			/* Just skip (strip) the colour codes */
Darkvater@5165: 			str += len;
Darkvater@5165: 		}
Darkvater@5165: 	}
Darkvater@5165: 	*dst = '\0';
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: /** Convert a given ASCII string to lowercase.
Darkvater@5165:  * NOTE: only support ASCII characters, no UTF8 fancy. As currently
Darkvater@5165:  * the function is only used to lowercase data-filenames if they are
Darkvater@5165:  * not found, this is sufficient. If more, or general functionality is
Darkvater@5165:  * needed, look to r7271 where it was removed because it was broken when
Darkvater@5165:  * using certain locales: eg in Turkish the uppercase 'I' was converted to
belugas@6420:  * '?', so just revert to the old functionality
belugas@6420:  * @param str string to convert */
Darkvater@5165: void strtolower(char *str)
Darkvater@5165: {
Darkvater@5165: 	for (; *str != '\0'; str++) *str = tolower(*str);
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: /**
Darkvater@5165:  * Only allow certain keys. You can define the filter to be used. This makes
Darkvater@5165:  *  sure no invalid keys can get into an editbox, like BELL.
Darkvater@5165:  * @param key character to be checked
Darkvater@5165:  * @param afilter the filter to use
Darkvater@5165:  * @return true or false depending if the character is printable/valid or not
Darkvater@5165:  */
Darkvater@5165: bool IsValidChar(WChar key, CharSetFilter afilter)
Darkvater@5165: {
Darkvater@5165: 	switch (afilter) {
Darkvater@5165: 		case CS_ALPHANUMERAL: return IsPrintable(key);
Darkvater@5165: 		case CS_NUMERAL:      return (key >= '0' && key <= '9');
Darkvater@5165: 		case CS_ALPHA:        return IsPrintable(key) && !(key >= '0' && key <= '9');
Darkvater@5165: 	}
Darkvater@5165: 
Darkvater@5165: 	return false;
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: #ifdef WIN32
glx@8181: /* Since version 3.14, MinGW Runtime has snprintf() and vsnprintf() conform to C99 but it's not the case for older versions */
glx@8181: #if (__MINGW32_MAJOR_VERSION < 3) || ((__MINGW32_MAJOR_VERSION == 3) && (__MINGW32_MINOR_VERSION < 14))
Darkvater@5165: int CDECL snprintf(char *str, size_t size, const char *format, ...)
Darkvater@5165: {
Darkvater@5165: 	va_list ap;
Darkvater@5165: 	int ret;
Darkvater@5165: 
Darkvater@5165: 	va_start(ap, format);
Darkvater@5165: 	ret = vsnprintf(str, size, format, ap);
Darkvater@5165: 	va_end(ap);
Darkvater@5165: 	return ret;
Darkvater@5165: }
glx@8181: #endif /* MinGW Runtime < 3.14 */
Darkvater@5165: 
Darkvater@5165: #ifdef _MSC_VER
Darkvater@5630: /* *nprintf broken, not POSIX compliant, MSDN description
Darkvater@5630:  * - If len < count, then len characters are stored in buffer, a null-terminator is appended, and len is returned.
Darkvater@5630:  * - If len = count, then len characters are stored in buffer, no null-terminator is appended, and len is returned.
Darkvater@5630:  * - If len > count, then count characters are stored in buffer, no null-terminator is appended, and a negative value is returned
Darkvater@5630:  */
Darkvater@5165: int CDECL vsnprintf(char *str, size_t size, const char *format, va_list ap)
Darkvater@5165: {
Darkvater@5165: 	int ret;
Darkvater@5165: 	ret = _vsnprintf(str, size, format, ap);
Darkvater@5631: 	if (ret < 0 || ret == size) str[size - 1] = '\0';
Darkvater@5165: 	return ret;
Darkvater@5165: }
Darkvater@5165: #endif /* _MSC_VER */
Darkvater@5165: 
Darkvater@5165: #endif /* WIN32 */
Darkvater@5165: 
rubidium@10299: /**
rubidium@10299:  * Safer implementation of snprintf; same as snprintf except:
rubidium@10299:  * - last instead of size, i.e. replace sizeof with lastof.
rubidium@10299:  * - return gives the amount of characters added, not what it would add.
rubidium@10299:  * @param str    buffer to write to up to last
rubidium@10299:  * @param last   last character we may write to
rubidium@10299:  * @param format the formatting (see snprintf)
rubidium@10299:  * @return the number of added characters
rubidium@10299:  */
rubidium@10299: int CDECL seprintf(char *str, const char *last, const char *format, ...)
rubidium@10299: {
rubidium@10299: 	va_list ap;
rubidium@10299: 
rubidium@10299: 	va_start(ap, format);
rubidium@10299: 	int ret = vseprintf(str, last, format, ap);
rubidium@10299: 	va_end(ap);
rubidium@10299: 	return ret;
rubidium@10299: }
rubidium@10299: 
Darkvater@5165: 
Darkvater@5634: /** Convert the md5sum to a hexadecimal string representation
Darkvater@5634:  * @param buf buffer to put the md5sum into
Darkvater@5634:  * @param last last character of buffer (usually lastof(buf))
Darkvater@5634:  * @param md5sum the md5sum itself
Darkvater@5634:  * @return a pointer to the next character after the md5sum */
Darkvater@5634: char *md5sumToString(char *buf, const char *last, const uint8 md5sum[16])
Darkvater@5634: {
Darkvater@5634: 	char *p = buf;
Darkvater@5634: 
Darkvater@5634: 	for (uint i = 0; i < 16; i++) {
rubidium@10299: 		p += seprintf(p, last, "%02X", md5sum[i]);
Darkvater@5634: 	}
Darkvater@5634: 
Darkvater@5634: 	return p;
Darkvater@5634: }
Darkvater@5634: 
Darkvater@5634: 
Darkvater@5165: /* UTF-8 handling routines */
Darkvater@5165: 
Darkvater@5165: 
Darkvater@5165: /* Decode and consume the next UTF-8 encoded character
Darkvater@5165:  * @param c Buffer to place decoded character.
Darkvater@5165:  * @param s Character stream to retrieve character from.
Darkvater@5165:  * @return Number of characters in the sequence.
Darkvater@5165:  */
Darkvater@5165: size_t Utf8Decode(WChar *c, const char *s)
Darkvater@5165: {
Darkvater@5165: 	assert(c != NULL);
Darkvater@5165: 
skidd13@7928: 	if (!HasBit(s[0], 7)) {
Darkvater@5165: 		/* Single byte character: 0xxxxxxx */
Darkvater@5165: 		*c = s[0];
Darkvater@5165: 		return 1;
Darkvater@5165: 	} else if (GB(s[0], 5, 3) == 6) {
Darkvater@5165: 		if (IsUtf8Part(s[1])) {
Darkvater@5165: 			/* Double byte character: 110xxxxx 10xxxxxx */
Darkvater@5165: 			*c = GB(s[0], 0, 5) << 6 | GB(s[1], 0, 6);
Darkvater@5165: 			if (*c >= 0x80) return 2;
Darkvater@5165: 		}
Darkvater@5165: 	} else if (GB(s[0], 4, 4) == 14) {
Darkvater@5165: 		if (IsUtf8Part(s[1]) && IsUtf8Part(s[2])) {
Darkvater@5165: 			/* Triple byte character: 1110xxxx 10xxxxxx 10xxxxxx */
Darkvater@5165: 			*c = GB(s[0], 0, 4) << 12 | GB(s[1], 0, 6) << 6 | GB(s[2], 0, 6);
Darkvater@5165: 			if (*c >= 0x800) return 3;
Darkvater@5165: 		}
Darkvater@5165: 	} else if (GB(s[0], 3, 5) == 30) {
Darkvater@5165: 		if (IsUtf8Part(s[1]) && IsUtf8Part(s[2]) && IsUtf8Part(s[3])) {
Darkvater@5165: 			/* 4 byte character: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
Darkvater@5165: 			*c = GB(s[0], 0, 3) << 18 | GB(s[1], 0, 6) << 12 | GB(s[2], 0, 6) << 6 | GB(s[3], 0, 6);
Darkvater@5165: 			if (*c >= 0x10000 && *c <= 0x10FFFF) return 4;
Darkvater@5165: 		}
Darkvater@5165: 	}
Darkvater@5165: 
Darkvater@5380: 	//DEBUG(misc, 1, "[utf8] invalid UTF-8 sequence");
Darkvater@5165: 	*c = '?';
Darkvater@5165: 	return 1;
Darkvater@5165: }
Darkvater@5165: 
Darkvater@5165: 
Darkvater@5165: /* Encode a unicode character and place it in the buffer
Darkvater@5165:  * @param buf Buffer to place character.
Darkvater@5165:  * @param c   Unicode character to encode.
Darkvater@5165:  * @return Number of characters in the encoded sequence.
Darkvater@5165:  */
Darkvater@5165: size_t Utf8Encode(char *buf, WChar c)
Darkvater@5165: {
Darkvater@5165: 	if (c < 0x80) {
Darkvater@5165: 		*buf = c;
Darkvater@5165: 		return 1;
Darkvater@5165: 	} else if (c < 0x800) {
Darkvater@5165: 		*buf++ = 0xC0 + GB(c,  6, 5);
Darkvater@5165: 		*buf   = 0x80 + GB(c,  0, 6);
Darkvater@5165: 		return 2;
Darkvater@5165: 	} else if (c < 0x10000) {
Darkvater@5165: 		*buf++ = 0xE0 + GB(c, 12, 4);
Darkvater@5165: 		*buf++ = 0x80 + GB(c,  6, 6);
Darkvater@5165: 		*buf   = 0x80 + GB(c,  0, 6);
Darkvater@5165: 		return 3;
Darkvater@5165: 	} else if (c < 0x110000) {
Darkvater@5165: 		*buf++ = 0xF0 + GB(c, 18, 3);
Darkvater@5165: 		*buf++ = 0x80 + GB(c, 12, 6);
Darkvater@5165: 		*buf++ = 0x80 + GB(c,  6, 6);
Darkvater@5165: 		*buf   = 0x80 + GB(c,  0, 6);
Darkvater@5165: 		return 4;
Darkvater@5165: 	}
Darkvater@5165: 
Darkvater@5380: 	//DEBUG(misc, 1, "[utf8] can't UTF-8 encode value 0x%X", c);
Darkvater@5165: 	*buf = '?';
Darkvater@5165: 	return 1;
Darkvater@5165: }
Darkvater@6215: 
Darkvater@6215: /**
Darkvater@6215:  * Properly terminate an UTF8 string to some maximum length
Darkvater@6215:  * @param s string to check if it needs additional trimming
Darkvater@6215:  * @param maxlen the maximum length the buffer can have.
Darkvater@6215:  * @return the new length in bytes of the string (eg. strlen(new_string))
Darkvater@6215:  * @NOTE maxlen is the string length _INCLUDING_ the terminating '\0'
Darkvater@6215:  */
Darkvater@6215: size_t Utf8TrimString(char *s, size_t maxlen)
Darkvater@6215: {
Darkvater@6215: 	size_t length = 0;
Darkvater@6215: 
Darkvater@6215: 	for (const char *ptr = strchr(s, '\0'); *s != '\0';) {
Darkvater@6215: 		size_t len = Utf8EncodedCharLen(*s);
Darkvater@6274: 		/* Silently ignore invalid UTF8 sequences, our only concern trimming */
Darkvater@6274: 		if (len == 0) len = 1;
Darkvater@6215: 
Darkvater@6215: 		/* Take care when a hard cutoff was made for the string and
Darkvater@6215: 		 * the last UTF8 sequence is invalid */
Darkvater@6215: 		if (length + len >= maxlen || (s + len > ptr)) break;
Darkvater@6215: 		s += len;
Darkvater@6215: 		length += len;
Darkvater@6215: 	}
Darkvater@6215: 
Darkvater@6215: 	*s = '\0';
Darkvater@6215: 	return length;
tron@6218: }
glx@9997: 
glx@9997: #ifndef _GNU_SOURCE
glx@9997: #include "core/math_func.hpp"
glx@9997: char *strndup(const char *s, size_t len)
glx@9997: {
glx@9997: 	len = min(strlen(s), len);
glx@9997: 	char *tmp = CallocT<char>(len + 1);
glx@9997: 	memcpy(tmp, s, len);
glx@9997: 	return tmp;
glx@9997: }
glx@9997: #endif /* !_GNU_SOURCE */