/* Implementation of extension methods to standard classes Copyright (C) 2003 Free Software Foundation, Inc. Written by: Richard Frith-Macdonald This file is part of the GNUstep Base Library. This library is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public License for more details. You should have received a copy of the GNU Library General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111 USA. */ #include "config.h" #include #include #include "GNUstepBase/GSCategories.h" #include "GNUstepBase/GSLock.h" /** * Extension methods for the NSCalendarDate class */ @implementation NSCalendarDate (GSCategories) /** * The ISO standard week of the year is based on the first week of the * year being that week (starting on monday) for which the thursday * is on or after the first of january.
* This has the effect that, if january first is a friday, saturday or * sunday, the days of that week (up to and including the sunday) are * considered to be in week 53 of the preceeding year. Similarly if the * last day of the year is a monday tuesday or wednesday, these days are * part of week 1 of the next year. */ - (int) weekOfYear { int dayOfWeek = [self dayOfWeek]; int dayOfYear; /* * Whether a week is considered to be in a year or not depends on its * thursday ... so find thursday for the receivers week. * NB. this may result in a date which is not in the same year as the * receiver. */ if (dayOfWeek != 4) { CREATE_AUTORELEASE_POOL(arp); NSCalendarDate *thursday; /* * A week starts on monday ... so adjust from 0 to 7 so that a * sunday is counted as the last day of the week. */ if (dayOfWeek == 0) { dayOfWeek = 7; } thursday = [self dateByAddingYears: 0 months: 0 days: 4 - dayOfWeek hours: 0 minutes: 0 seconds: 0]; dayOfYear = [thursday dayOfYear]; RELEASE(arp); } else { dayOfYear = [self dayOfYear]; } /* * Round up to a week boundary, so that when we divide by seven we * get a result in the range 1 to 53 as mandated by the ISO standard. */ dayOfYear += (7 - dayOfYear % 7); return dayOfYear / 7; } @end /** * Extension methods for the NSData class */ @implementation NSData (GSCategories) /** * Returns an NSString object containing an ASCII hexadecimal representation * of the receiver. This means that the returned object will contain * exactly twice as many characters as there are bytes as the receiver, * as each byte in the receiver is represented by two hexadecimal digits.
* The high order four bits of each byte is encoded before the low * order four bits. Capital letters 'A' to 'F' are used to represent * values from 10 to 15.
* If you need the hexadecimal representation as raw byte data, use code * like - * * hexData = [[sourceData hexadecimalRepresentation] * dataUsingEncoding: NSASCIIStringEncoding]; * */ - (NSString*) hexadecimalRepresentation { static const char *hexChars = "0123456789ABCDEF"; unsigned slen = [self length]; unsigned dlen = slen * 2; const unsigned char *src = (const unsigned char *)[self bytes]; char *dst = (char*)NSZoneMalloc(NSDefaultMallocZone(), dlen); unsigned spos = 0; unsigned dpos = 0; NSData *data; NSString *string; while (spos < slen) { unsigned char c = src[spos++]; dst[dpos++] = hexChars[(c >> 4) & 0x0f]; dst[dpos++] = hexChars[c & 0x0f]; } data = [NSData allocWithZone: NSDefaultMallocZone()]; data = [data initWithBytesNoCopy: dst length: dlen freeWhenDone: YES]; string = [[NSString alloc] initWithData: data encoding: NSASCIIStringEncoding]; RELEASE(data); return AUTORELEASE(string); } /** * Initialises the receiver with the supplied string data which contains * a hexadecimal coding of the bytes. The parsing of the string is * fairly tolerant, ignoring whitespace and permitting both upper and * lower case hexadecimal digits (the -hexadecimalRepresentation method * produces a string using only uppercase digits with no white spaqce).
* If the string does not contain one or more pairs of hexadecimal digits * then an exception is raised. */ - (id) initWithHexadecimalRepresentation: (NSString*)string { CREATE_AUTORELEASE_POOL(arp); NSData *d; const char *src; const char *end; unsigned char *dst; unsigned int pos = 0; unsigned char byte = 0; BOOL high = NO; d = [string dataUsingEncoding: NSASCIIStringEncoding allowLossyConversion: YES]; src = (const char*)[d bytes]; end = src + [d length]; dst = NSZoneMalloc(NSDefaultMallocZone(), [d length]/2 + 1); while (src < end) { char c = *src++; unsigned char v; if (isspace(c)) { continue; } if (c >= '0' && c <= '9') { v = c - '0'; } else if (c >= 'A' && c <= 'F') { v = c - 'A' + 10; } else if (c >= 'a' && c <= 'f') { v = c - 'a' + 10; } else { pos = 0; break; } if (high == NO) { byte = v << 4; high = YES; } else { byte |= v; high = NO; dst[pos++] = byte; } } if (pos > 0 && high == NO) { self = [self initWithBytes: dst length: pos]; } else { DESTROY(self); } NSZoneFree(NSDefaultMallocZone(), dst); RELEASE(arp); if (self == nil) { [NSException raise: NSInvalidArgumentException format: @"%@: invalid hexadeciaml string data", NSStringFromSelector(_cmd)]; } return self; } struct MD5Context { unsigned long buf[4]; unsigned long bits[2]; unsigned char in[64]; }; static void MD5Init (struct MD5Context *context); static void MD5Update (struct MD5Context *context, unsigned char const *buf, unsigned len); static void MD5Final (unsigned char digest[16], struct MD5Context *context); static void MD5Transform (unsigned long buf[4], unsigned long const in[16]); /* * This code implements the MD5 message-digest algorithm. * The algorithm is due to Ron Rivest. This code was * written by Colin Plumb in 1993, no copyright is claimed. * This code is in the public domain; do with it what you wish. * * Equivalent code is available from RSA Data Security, Inc. * This code has been tested against that, and is equivalent, * except that you don't need to include two pages of legalese * with every copy. * * To compute the message digest of a chunk of bytes, declare an * MD5Context structure, pass it to MD5Init, call MD5Update as * needed on buffers full of bytes, and then call MD5Final, which * will fill a supplied 16-byte array with the digest. */ /* * Ensure data is little-endian */ static void littleEndian (void *buf, unsigned longs) { unsigned long *ptr = (unsigned long*)buf; do { *ptr = NSSwapHostLongToLittle(*ptr); ptr++; } while (--longs); } /* * Start MD5 accumulation. Set bit count to 0 and buffer to mysterious * initialization constants. */ static void MD5Init (struct MD5Context *ctx) { ctx->buf[0] = 0x67452301; ctx->buf[1] = 0xefcdab89; ctx->buf[2] = 0x98badcfe; ctx->buf[3] = 0x10325476; ctx->bits[0] = 0; ctx->bits[1] = 0; } /* * Update context to reflect the concatenation of another buffer full * of bytes. */ static void MD5Update (struct MD5Context *ctx, unsigned char const *buf, unsigned len) { unsigned long t; /* Update bitcount */ t = ctx->bits[0]; if ((ctx->bits[0] = t + ((unsigned long) len << 3)) < t) ctx->bits[1]++; /* Carry from low to high */ ctx->bits[1] += len >> 29; t = (t >> 3) & 0x3f; /* Bytes already in shsInfo->data */ /* Handle any leading odd-sized chunks */ if (t) { unsigned char *p = (unsigned char *) ctx->in + t; t = 64 - t; if (len < t) { memcpy (p, buf, len); return; } memcpy (p, buf, t); littleEndian (ctx->in, 16); MD5Transform (ctx->buf, (unsigned long *) ctx->in); buf += t; len -= t; } /* Process data in 64-byte chunks */ while (len >= 64) { memcpy (ctx->in, buf, 64); littleEndian (ctx->in, 16); MD5Transform (ctx->buf, (unsigned long *) ctx->in); buf += 64; len -= 64; } /* Handle any remaining bytes of data. */ memcpy (ctx->in, buf, len); } /* * Final wrapup - pad to 64-byte boundary with the bit pattern * 1 0* (64-bit count of bits processed, MSB-first) */ static void MD5Final (unsigned char digest[16], struct MD5Context *ctx) { unsigned count; unsigned char *p; /* Compute number of bytes mod 64 */ count = (ctx->bits[0] >> 3) & 0x3F; /* Set the first char of padding to 0x80. This is safe since there is always at least one byte free */ p = ctx->in + count; *p++ = 0x80; /* Bytes of padding needed to make 64 bytes */ count = 64 - 1 - count; /* Pad out to 56 mod 64 */ if (count < 8) { /* Two lots of padding: Pad the first block to 64 bytes */ memset (p, 0, count); littleEndian (ctx->in, 16); MD5Transform (ctx->buf, (unsigned long *) ctx->in); /* Now fill the next block with 56 bytes */ memset (ctx->in, 0, 56); } else { /* Pad block to 56 bytes */ memset (p, 0, count - 8); } littleEndian (ctx->in, 14); /* Append length in bits and transform */ ((unsigned long *) ctx->in)[14] = ctx->bits[0]; ((unsigned long *) ctx->in)[15] = ctx->bits[1]; MD5Transform (ctx->buf, (unsigned long *) ctx->in); littleEndian ((unsigned char *) ctx->buf, 4); memcpy (digest, ctx->buf, 16); memset (ctx, 0, sizeof (ctx)); /* In case it's sensitive */ } /* The four core functions - F1 is optimized somewhat */ /* #define F1(x, y, z) (x & y | ~x & z) */ #define F1(x, y, z) (z ^ (x & (y ^ z))) #define F2(x, y, z) F1(z, x, y) #define F3(x, y, z) (x ^ y ^ z) #define F4(x, y, z) (y ^ (x | ~z)) /* This is the central step in the MD5 algorithm. */ #define MD5STEP(f, w, x, y, z, data, s) \ ( w += f(x, y, z) + data, w = w<>(32-s), w += x ) /* * The core of the MD5 algorithm, this alters an existing MD5 hash to * reflect the addition of 16 longwords of new data. MD5Update blocks * the data and converts bytes into longwords for this routine. */ static void MD5Transform (unsigned long buf[4], unsigned long const in[16]) { register unsigned long a, b, c, d; a = buf[0]; b = buf[1]; c = buf[2]; d = buf[3]; MD5STEP (F1, a, b, c, d, in[0] + 0xd76aa478, 7); MD5STEP (F1, d, a, b, c, in[1] + 0xe8c7b756, 12); MD5STEP (F1, c, d, a, b, in[2] + 0x242070db, 17); MD5STEP (F1, b, c, d, a, in[3] + 0xc1bdceee, 22); MD5STEP (F1, a, b, c, d, in[4] + 0xf57c0faf, 7); MD5STEP (F1, d, a, b, c, in[5] + 0x4787c62a, 12); MD5STEP (F1, c, d, a, b, in[6] + 0xa8304613, 17); MD5STEP (F1, b, c, d, a, in[7] + 0xfd469501, 22); MD5STEP (F1, a, b, c, d, in[8] + 0x698098d8, 7); MD5STEP (F1, d, a, b, c, in[9] + 0x8b44f7af, 12); MD5STEP (F1, c, d, a, b, in[10] + 0xffff5bb1, 17); MD5STEP (F1, b, c, d, a, in[11] + 0x895cd7be, 22); MD5STEP (F1, a, b, c, d, in[12] + 0x6b901122, 7); MD5STEP (F1, d, a, b, c, in[13] + 0xfd987193, 12); MD5STEP (F1, c, d, a, b, in[14] + 0xa679438e, 17); MD5STEP (F1, b, c, d, a, in[15] + 0x49b40821, 22); MD5STEP (F2, a, b, c, d, in[1] + 0xf61e2562, 5); MD5STEP (F2, d, a, b, c, in[6] + 0xc040b340, 9); MD5STEP (F2, c, d, a, b, in[11] + 0x265e5a51, 14); MD5STEP (F2, b, c, d, a, in[0] + 0xe9b6c7aa, 20); MD5STEP (F2, a, b, c, d, in[5] + 0xd62f105d, 5); MD5STEP (F2, d, a, b, c, in[10] + 0x02441453, 9); MD5STEP (F2, c, d, a, b, in[15] + 0xd8a1e681, 14); MD5STEP (F2, b, c, d, a, in[4] + 0xe7d3fbc8, 20); MD5STEP (F2, a, b, c, d, in[9] + 0x21e1cde6, 5); MD5STEP (F2, d, a, b, c, in[14] + 0xc33707d6, 9); MD5STEP (F2, c, d, a, b, in[3] + 0xf4d50d87, 14); MD5STEP (F2, b, c, d, a, in[8] + 0x455a14ed, 20); MD5STEP (F2, a, b, c, d, in[13] + 0xa9e3e905, 5); MD5STEP (F2, d, a, b, c, in[2] + 0xfcefa3f8, 9); MD5STEP (F2, c, d, a, b, in[7] + 0x676f02d9, 14); MD5STEP (F2, b, c, d, a, in[12] + 0x8d2a4c8a, 20); MD5STEP (F3, a, b, c, d, in[5] + 0xfffa3942, 4); MD5STEP (F3, d, a, b, c, in[8] + 0x8771f681, 11); MD5STEP (F3, c, d, a, b, in[11] + 0x6d9d6122, 16); MD5STEP (F3, b, c, d, a, in[14] + 0xfde5380c, 23); MD5STEP (F3, a, b, c, d, in[1] + 0xa4beea44, 4); MD5STEP (F3, d, a, b, c, in[4] + 0x4bdecfa9, 11); MD5STEP (F3, c, d, a, b, in[7] + 0xf6bb4b60, 16); MD5STEP (F3, b, c, d, a, in[10] + 0xbebfbc70, 23); MD5STEP (F3, a, b, c, d, in[13] + 0x289b7ec6, 4); MD5STEP (F3, d, a, b, c, in[0] + 0xeaa127fa, 11); MD5STEP (F3, c, d, a, b, in[3] + 0xd4ef3085, 16); MD5STEP (F3, b, c, d, a, in[6] + 0x04881d05, 23); MD5STEP (F3, a, b, c, d, in[9] + 0xd9d4d039, 4); MD5STEP (F3, d, a, b, c, in[12] + 0xe6db99e5, 11); MD5STEP (F3, c, d, a, b, in[15] + 0x1fa27cf8, 16); MD5STEP (F3, b, c, d, a, in[2] + 0xc4ac5665, 23); MD5STEP (F4, a, b, c, d, in[0] + 0xf4292244, 6); MD5STEP (F4, d, a, b, c, in[7] + 0x432aff97, 10); MD5STEP (F4, c, d, a, b, in[14] + 0xab9423a7, 15); MD5STEP (F4, b, c, d, a, in[5] + 0xfc93a039, 21); MD5STEP (F4, a, b, c, d, in[12] + 0x655b59c3, 6); MD5STEP (F4, d, a, b, c, in[3] + 0x8f0ccc92, 10); MD5STEP (F4, c, d, a, b, in[10] + 0xffeff47d, 15); MD5STEP (F4, b, c, d, a, in[1] + 0x85845dd1, 21); MD5STEP (F4, a, b, c, d, in[8] + 0x6fa87e4f, 6); MD5STEP (F4, d, a, b, c, in[15] + 0xfe2ce6e0, 10); MD5STEP (F4, c, d, a, b, in[6] + 0xa3014314, 15); MD5STEP (F4, b, c, d, a, in[13] + 0x4e0811a1, 21); MD5STEP (F4, a, b, c, d, in[4] + 0xf7537e82, 6); MD5STEP (F4, d, a, b, c, in[11] + 0xbd3af235, 10); MD5STEP (F4, c, d, a, b, in[2] + 0x2ad7d2bb, 15); MD5STEP (F4, b, c, d, a, in[9] + 0xeb86d391, 21); buf[0] += a; buf[1] += b; buf[2] += c; buf[3] += d; } /** * Creates an MD5 digest of the information stored in the receiver and * returns it as an autoreleased 16 byte NSData object.
* If you need to produce a digest of string information, you need to * decide what character encoding is to be used and convert your string * to a data object of that encoding type first using the * [NSString-dataUsingEncoding:] method - * * myDigest = [[myString dataUsingEncoding: NSUTF8StringEncoding] md5Digest]; * * If you need to use the digest in a human readable form, you will * probably want it to be seen as 32 hexadecimal digits, and can do that * using the -hexadecimalRepresentation method. */ - (NSData*) md5Digest { struct MD5Context ctx; unsigned char digest[16]; MD5Init(&ctx); MD5Update(&ctx, [self bytes], [self length]); MD5Final(digest, &ctx); return [NSData dataWithBytes: digest length: 16]; } /** * Decodes the source data from uuencoded and return the result.
* Returns the encoded file name in namePtr if it is not null. * Returns the encoded file mode in modePtr if it is not null. */ - (BOOL) uudecodeInto: (NSMutableData*)decoded name: (NSString**)namePtr mode: (int*)modePtr { const unsigned char *bytes = (const unsigned char*)[self bytes]; unsigned length = [self length]; unsigned decLength = [decoded length]; unsigned pos = 0; NSString *name = nil; if (namePtr != 0) { *namePtr = nil; } if (modePtr != 0) { *modePtr = 0; } #define DEC(c) (((c) - ' ') & 077) for (pos = 0; pos < length; pos++) { if (bytes[pos] == '\n') { if (name != nil) { unsigned i = 0; int lineLength; unsigned char *decPtr; lineLength = DEC(bytes[i++]); if (lineLength <= 0) { break; // Got line length zero or less. } [decoded setLength: decLength + lineLength]; decPtr = [decoded mutableBytes]; while (lineLength > 0) { unsigned char tmp[4]; int c; /* * In case the data is corrupt, we need to copy into * a temporary buffer avoiding buffer overrun in the * main buffer. */ tmp[0] = bytes[i++]; if (i < pos) { tmp[1] = bytes[i++]; if (i < pos) { tmp[2] = bytes[i++]; if (i < pos) { tmp[3] = bytes[i++]; } else { tmp[3] = 0; } } else { tmp[2] = 0; tmp[3] = 0; } } else { tmp[1] = 0; tmp[2] = 0; tmp[3] = 0; } if (lineLength >= 1) { c = DEC(tmp[0]) << 2 | DEC(tmp[1]) >> 4; decPtr[decLength++] = (unsigned char)c; } if (lineLength >= 2) { c = DEC(tmp[1]) << 4 | DEC(tmp[2]) >> 2; decPtr[decLength++] = (unsigned char)c; } if (lineLength >= 3) { c = DEC(tmp[2]) << 6 | DEC(tmp[3]); decPtr[decLength++] = (unsigned char)c; } lineLength -= 3; } } else if (pos > 6 && strncmp(bytes, "begin ", 6) == 0) { unsigned off = 6; unsigned end = pos; int mode = 0; NSData *d; if (end > off && bytes[end-1] == '\r') { end--; } while (off < end && bytes[off] >= '0' && bytes[off] <= '7') { mode *= 8; mode += bytes[off] - '0'; off++; } if (modePtr != 0) { *modePtr = mode; } while (off < end && bytes[off] == ' ') { off++; } d = [NSData dataWithBytes: &bytes[off] length: end - off]; name = [[NSString alloc] initWithData: d encoding: NSASCIIStringEncoding]; AUTORELEASE(name); if (namePtr != 0) { *namePtr = name; } } pos++; bytes += pos; length -= pos; } } if (name == nil) { return NO; } return YES; } /** * Encode the source data to uuencoded.
* Uses the supplied name as the filename in the encoded data, * and says that the file mode is as specified.
* If no name is supplied, uses untitled as the name. */ - (BOOL) uuencodeInto: (NSMutableData*)encoded name: (NSString*)name mode: (int)mode { const unsigned char *bytes = (const unsigned char*)[self bytes]; int length = [self length]; unsigned char buf[64]; unsigned i; name = [name stringByTrimmingSpaces]; if ([name length] == 0) { name = @"untitled"; } /* * The header is a line of the form 'begin mode filename' */ sprintf(buf, "begin %03o ", mode); [encoded appendBytes: buf length: strlen(buf)]; [encoded appendData: [name dataUsingEncoding: NSASCIIStringEncoding]]; [encoded appendBytes: "\n" length: 1]; #define ENC(c) ((c) > 0 ? ((c) & 077) + ' ': '`') while (length > 0) { int count; unsigned pos; /* * We want up to 45 bytes in a line ... and we record the * number of bytes as the initial output character. */ count = length; if (count > 45) { count = 45; } i = 0; buf[i++] = ENC(count); /* * Now we encode the actual data for the line. */ for (pos = 0; count > 0; count -= 3) { unsigned char tmp[3]; int c; /* * Copy data into a temporary buffer ensuring we don't * overrun the end of the original buffer risking access * violation. */ tmp[0] = bytes[pos++]; if (pos < length) { tmp[1] = bytes[pos++]; if (pos < length) { tmp[2] = bytes[pos++]; } else { tmp[2] = 0; } } else { tmp[1] = 0; tmp[2] = 0; } c = tmp[0] >> 2; buf[i++] = ENC(c); c = ((tmp[0] << 4) & 060) | ((tmp[1] >> 4) & 017); buf[i++] = ENC(c); c = ((tmp[1] << 2) & 074) | ((tmp[2] >> 6) & 03); buf[i++] = ENC(c); c = tmp[2] & 077; buf[i++] = ENC(c); } bytes += pos; length -= pos; buf[i++] = '\n'; [encoded appendBytes: buf length: i]; } /* * Encode a line of length zero followed by 'end' as the terminator. */ [encoded appendBytes: "`\n" length: 4]; [encoded appendBytes: "end\n" length: 4]; return YES; } @end /** * GNUstep specific (non-standard) additions to the NSNumber class. */ @implementation NSNumber(GSCategories) + (NSValue*) valueFromString: (NSString*)string { /* FIXME: implement this better */ const char *str; str = [string cString]; if (strchr(str, '.') >= 0 || strchr(str, 'e') >= 0 || strchr(str, 'E') >= 0) return [NSNumber numberWithDouble: atof(str)]; else if (strchr(str, '-') >= 0) return [NSNumber numberWithInt: atoi(str)]; else return [NSNumber numberWithUnsignedInt: atoi(str)]; return [NSNumber numberWithInt: 0]; } @end /** * Extension methods for the NSObject class */ @implementation NSObject (GSCategories) - (id) notImplemented: (SEL)aSel { [NSException raise: NSGenericException format: @"method %s not implemented in %s(%s)", aSel ? GSNameFromSelector(aSel) : "(null)", GSClassNameFromObject(self), GSObjCIsInstance(self) ? "instance" : "class"]; return nil; } - (id) shouldNotImplement: (SEL)aSel { [NSException raise: NSGenericException format: @"%s(%s) should not implement %s", GSClassNameFromObject(self), GSObjCIsInstance(self) ? "instance" : "class", aSel ? GSNameFromSelector(aSel) : "(null)"]; return nil; } - (id) subclassResponsibility: (SEL)aSel { [NSException raise: NSGenericException format: @"subclass %s(%s) should override %s", GSClassNameFromObject(self), GSObjCIsInstance(self) ? "instance" : "class", aSel ? GSNameFromSelector(aSel) : "(null)"]; return nil; } /** * WARNING: The -compare: method for NSObject is deprecated * due to subclasses declaring the same selector with * conflicting signatures. * Comparision of arbitrary objects is not just meaningless * but also dangerous as most concrete implementations * expect comparable objects as arguments often accessing * instance variables directly. * This method will be removed in a future release. */ - (NSComparisonResult) compare: (id)anObject { NSLog(@"WARNING: The -compare: method for NSObject is deprecated."); if (anObject == self) { return NSOrderedSame; } if (anObject == nil) { [NSException raise: NSInvalidArgumentException format: @"nil argument for compare:"]; } if ([self isEqual: anObject]) { return NSOrderedSame; } /* * Ordering objects by their address is pretty useless, * so subclasses should override this is some useful way. */ if (self > anObject) { return NSOrderedDescending; } else { return NSOrderedAscending; } } @end /** * GNUstep specific (non-standard) additions to the NSString class. */ @implementation NSString (GSCategories) /** * Returns an autoreleased string initialized with -initWithFormat:arguments:. */ + (id) stringWithFormat: (NSString*)format arguments: (va_list)argList { return AUTORELEASE([[self allocWithZone: NSDefaultMallocZone()] initWithFormat: format arguments: argList]); } /** * Returns a string formed by removing the prefix string from the * receiver. Raises an exception if the prefix is not present. */ - (NSString*) stringByDeletingPrefix: (NSString*)prefix { NSCAssert2([self hasPrefix: prefix], @"'%@' does not have the prefix '%@'", self, prefix); return [self substringFromIndex: [prefix length]]; } /** * Returns a string formed by removing the suffix string from the * receiver. Raises an exception if the suffix is not present. */ - (NSString*) stringByDeletingSuffix: (NSString*)suffix { NSCAssert2([self hasSuffix: suffix], @"'%@' does not have the suffix '%@'", self, suffix); return [self substringToIndex: ([self length] - [suffix length])]; } /** * Returns a string formed by removing leading white space from the * receiver. */ - (NSString*) stringByTrimmingLeadSpaces { unsigned length = [self length]; if (length > 0) { unsigned start = 0; unichar (*caiImp)(NSString*, SEL, unsigned int); SEL caiSel = @selector(characterAtIndex:); caiImp = (unichar (*)())[self methodForSelector: caiSel]; while (start < length && isspace((*caiImp)(self, caiSel, start))) { start++; } if (start > 0) { return [self substringFromIndex: start]; } } return self; } /** * Returns a string formed by removing trailing white space from the * receiver. */ - (NSString*) stringByTrimmingTailSpaces { unsigned length = [self length]; if (length > 0) { unsigned end = length; unichar (*caiImp)(NSString*, SEL, unsigned int); SEL caiSel = @selector(characterAtIndex:); caiImp = (unichar (*)())[self methodForSelector: caiSel]; while (end > 0) { if (!isspace((*caiImp)(self, caiSel, end - 1))) { break; } end--; } if (end < length) { return [self substringToIndex: end]; } } return self; } /** * Returns a string formed by removing both leading and trailing * white space from the receiver. */ - (NSString*) stringByTrimmingSpaces { unsigned length = [self length]; if (length > 0) { unsigned start = 0; unsigned end = length; unichar (*caiImp)(NSString*, SEL, unsigned int); SEL caiSel = @selector(characterAtIndex:); caiImp = (unichar (*)())[self methodForSelector: caiSel]; while (start < length && isspace((*caiImp)(self, caiSel, start))) { start++; } while (end > start) { if (!isspace((*caiImp)(self, caiSel, end - 1))) { break; } end--; } if (start > 0 || end < length) { if (start < end) { return [self substringFromRange: NSMakeRange(start, end - start)]; } else { return [NSString string]; } } } return self; } /** * Returns a string in which any (and all) occurrances of * replace in the receiver have been replaced with by. * Returns the receiver if replace * does not occur within the receiver. NB. an empty string is * not considered to exist within the receiver. */ - (NSString*) stringByReplacingString: (NSString*)replace withString: (NSString*)by { NSRange range = [self rangeOfString: replace]; if (range.length > 0) { NSMutableString *tmp = [self mutableCopy]; NSString *str; [tmp replaceString: replace withString: by]; str = AUTORELEASE([tmp copy]); RELEASE(tmp); return str; } else return self; } @end /** * GNUstep specific (non-standard) additions to the NSMutableString class. */ @implementation NSMutableString (GSCategories) /** * Removes the specified suffix from the string. Raises an exception * if the suffix is not present. */ - (void) deleteSuffix: (NSString*)suffix { NSCAssert2([self hasSuffix: suffix], @"'%@' does not have the suffix '%@'", self, suffix); [self deleteCharactersInRange: NSMakeRange([self length] - [suffix length], [suffix length])]; } /** * Removes the specified prefix from the string. Raises an exception * if the prefix is not present. */ - (void) deletePrefix: (NSString*)prefix { NSCAssert2([self hasPrefix: prefix], @"'%@' does not have the prefix '%@'", self, prefix); [self deleteCharactersInRange: NSMakeRange(0, [prefix length])]; } /** * Replaces all occurrances of the string replace with the string by * in the receiver.
* Has no effect if replace does not occur within the * receiver. NB. an empty string is not considered to exist within * the receiver.
* Calls - replaceOccurrencesOfString:withString:options:range: passing * zero for the options and a range from 0 with the length of the receiver. */ - (void) replaceString: (NSString*)replace withString: (NSString*)by { [self replaceOccurrencesOfString: replace withString: by options: 0 range: NSMakeRange(0, [self length])]; } /** * Removes all leading white space from the receiver. */ - (void) trimLeadSpaces { unsigned length = [self length]; if (length > 0) { unsigned start = 0; unichar (*caiImp)(NSString*, SEL, unsigned int); SEL caiSel = @selector(characterAtIndex:); caiImp = (unichar (*)())[self methodForSelector: caiSel]; while (start < length && isspace((*caiImp)(self, caiSel, start))) { start++; } if (start > 0) { [self deleteCharactersInRange: NSMakeRange(0, start)]; } } } /** * Removes all trailing white space from the receiver. */ - (void) trimTailSpaces { unsigned length = [self length]; if (length > 0) { unsigned end = length; unichar (*caiImp)(NSString*, SEL, unsigned int); SEL caiSel = @selector(characterAtIndex:); caiImp = (unichar (*)())[self methodForSelector: caiSel]; while (end > 0 && isspace((*caiImp)(self, caiSel, end - 1))) { end--; } if (end < length) { [self deleteCharactersInRange: NSMakeRange(end, length - end)]; } } } /** * Removes all leading or trailing white space from the receiver. */ - (void) trimSpaces { [self trimTailSpaces]; [self trimLeadSpaces]; } @end /** * GNUstep specific (non-standard) additions to the NSLock class. */ static GSLazyRecursiveLock *local_lock = nil; /* This class only exists to provide a thread safe mechanism to initialize local_lock as +initialize is called under a lock in ObjC runtimes. User code should resort to GS_INITIALIZED_LOCK(), which uses the +newLockAt: extension. */ @interface _GSLockInitializer : NSObject @end @implementation _GSLockInitializer + (void) initialize { if (local_lock == nil) { /* As we do not know whether creating custom locks may implicitly create other locks, we use a recursive lock. */ local_lock = [GSLazyRecursiveLock new]; } } @end GS_STATIC_INLINE id newLockAt(Class self, SEL _cmd, id *location) { if (location == 0) { [NSException raise: NSInvalidArgumentException format: @"'%@' called with nil location", NSStringFromSelector(_cmd)]; } if (*location == nil) { if (local_lock == nil) { [_GSLockInitializer class]; } [local_lock lock]; if (*location == nil) { *location = [[self alloc] init]; } [local_lock unlock]; } return *location; } @implementation NSLock (GSCategories) /** * Initializes the id pointed to by location * with a new instance of the receiver's class * in a thread safe manner, unless * it has been previously initialized. * Returns the contents pointed to by location. * The location is considered unintialized if it contains nil. *
* This method is used in the GS_INITIALIZED_LOCK macro * to initialize lock variables when it cannot be insured * that they can be initialized in a thread safe environment. * * NSLock *my_lock = nil; * * void function (void) * { * [GS_INITIALIZED_LOCK(my_lock, NSLock) lock]; * do_work (); * [my_lock unlock]; * } * * */ + (id) newLockAt: (id *)location { return newLockAt(self, _cmd, location); } @end @implementation NSRecursiveLock (GSCategories) /** * Initializes the id pointed to by location * with a new instance of the receiver's class * in a thread safe manner, unless * it has been previously initialized. * Returns the contents pointed to by location. * The location is considered unintialized if it contains nil. *
* This method is used in the GS_INITIALIZED_LOCK macro * to initialize lock variables when it cannot be insured * that they can be initialized in a thread safe environment. * * NSLock *my_lock = nil; * * void function (void) * { * [GS_INITIALIZED_LOCK(my_lock, NSRecursiveLock) lock]; * do_work (); * [my_lock unlock]; * } * * */ + (id) newLockAt: (id *)location { return newLockAt(self, _cmd, location); } @end