Context Navigation

source: trunk/src/helpers/stringh.c@ 144

Visit:

Last change on this file since 144 was 144, checked in by umoeller, 23 years ago
Misc minor fixes.
Property svn:eol-style set to `CRLF` Property svn:keywords set to `Author Date Id Revision`
File size: 58.1 KB

Line
1
2	/*
3	*@@sourcefile stringh.c:
4	* contains string/text helper functions. These are good for
5	* parsing/splitting strings and other stuff used throughout
6	* XWorkplace.
7	*
8	* Note that these functions are really a bunch of very mixed
9	* up string helpers, which you may or may not find helpful.
10	* If you're looking for string functions with memory
11	* management, look at xstring.c instead.
12	*
13	* Usage: All OS/2 programs.
14	*
15	* Function prefixes (new with V0.81):
16	* -- strh* string helper functions.
17	*
18	* Note: Version numbering in this file relates to XWorkplace version
19	* numbering.
20	*
21	*@@header "helpers\stringh.h"
22	*/
23
24	/*
25	* Copyright (C) 1997-2000 Ulrich Mller.
26	* Parts Copyright (C) 1991-1999 iMatix Corporation.
27	* This file is part of the "XWorkplace helpers" source package.
28	* This is free software; you can redistribute it and/or modify
29	* it under the terms of the GNU General Public License as published
30	* by the Free Software Foundation, in version 2 as it comes in the
31	* "COPYING" file of the XWorkplace main distribution.
32	* This program is distributed in the hope that it will be useful,
33	* but WITHOUT ANY WARRANTY; without even the implied warranty of
34	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
35	* GNU General Public License for more details.
36	*/
37
38	#define OS2EMX_PLAIN_CHAR
39	// this is needed for "os2emx.h"; if this is defined,
40	// emx will define PSZ as _signed_ char, otherwise
41	// as unsigned char
42
43	#define INCL_WINSHELLDATA
44	#define INCL_DOSERRORS
45	#include <os2.h>
46
47	#include <stdlib.h>
48	#include <stdio.h>
49	#include <string.h>
50	#include <ctype.h>
51	#include <math.h>
52
53	#include "setup.h" // code generation and debugging options
54
55	#define DONT_REPLACE_STRINGH_MALLOC
56	#include "helpers\stringh.h"
57	#include "helpers\xstring.h" // extended string helpers
58
59	#pragma hdrstop
60
61	/*
62	*@@category: Helpers\C helpers\String management
63	* See stringh.c and xstring.c.
64	*/
65
66	/*
67	*@@category: Helpers\C helpers\String management\C string helpers
68	* See stringh.c.
69	*/
70
71	#ifdef __DEBUG_MALLOC_ENABLED__
72
73	/*
74	*@@ strhStoreDebug:
75	* memory debug version of strhStore.
76	*
77	*@@added V0.9.16 (2001-12-08) [umoeller]
78	*/
79
80	APIRET strhStoreDebug(PSZ *ppszTarget,
81	PCSZ pcszSource,
82	PULONG pulLength, // out: length of new string (ptr can be NULL)
83	PCSZ pcszSourceFile,
84	unsigned long ulLine,
85	PCSZ pcszFunction)
86	{
87	ULONG ulLength = 0;
88
89	if (ppszTarget)
90	{
91	if (*ppszTarget)
92	free(*ppszTarget);
93
94	if ( (pcszSource)
95	&& (ulLength = strlen(pcszSource))
96	)
97	{
98	if (*ppszTarget = (PSZ)memdMalloc(ulLength + 1,
99	pcszSourceFile,
100	ulLine,
101	pcszFunction))
102	memcpy(*ppszTarget, pcszSource, ulLength + 1);
103	else
104	return (ERROR_NOT_ENOUGH_MEMORY);
105	}
106	else
107	*ppszTarget = NULL;
108	}
109
110	if (pulLength)
111	*pulLength = ulLength;
112
113	return (NO_ERROR);
114	}
115
116	#endif
117
118	/*
119	*@@ strhStore:
120	* stores a copy of the given string in the specified
121	* buffer. Uses strdup internally.
122	*
123	* If *ppszTarget != NULL, the previous string is freed
124	* and set to NULL.
125	* If pcszSource != NULL, a copy of it is stored in the
126	* buffer.
127	*
128	*@@added V0.9.16 (2001-12-06) [umoeller]
129	*/
130
131	APIRET strhStore(PSZ *ppszTarget,
132	PCSZ pcszSource,
133	PULONG pulLength) // out: length of new string (ptr can be NULL)
134	{
135	ULONG ulLength = 0;
136
137	if (ppszTarget)
138	{
139	if (*ppszTarget)
140	free(*ppszTarget);
141
142	if ( (pcszSource)
143	&& (ulLength = strlen(pcszSource))
144	)
145	{
146	if (*ppszTarget = (PSZ)malloc(ulLength + 1))
147	memcpy(*ppszTarget, pcszSource, ulLength + 1);
148	else
149	return (ERROR_NOT_ENOUGH_MEMORY);
150	}
151	else
152	*ppszTarget = NULL;
153	}
154	else
155	return (ERROR_INVALID_PARAMETER);
156
157	if (pulLength)
158	*pulLength = ulLength;
159
160	return (NO_ERROR);
161	}
162
163	/*
164	*@@ strhcpy:
165	* like strdup, but this one doesn't crash if string2 is NULL,
166	* but sets the first byte in string1 to \0 instead.
167	*
168	*@@added V0.9.14 (2001-08-01) [umoeller]
169	*/
170
171	PSZ strhcpy(PSZ string1, PCSZ string2)
172	{
173	if (string2)
174	return (strcpy(string1, string2));
175
176	*string1 = '\0';
177	return (string1);
178	}
179
180	#ifdef __DEBUG_MALLOC_ENABLED__
181
182	/*
183	*@@ strhdupDebug:
184	* memory debug version of strhdup.
185	*
186	*@@added V0.9.0 [umoeller]
187	*/
188
189	PSZ strhdupDebug(PCSZ pcszSource,
190	unsigned long *pulLength,
191	PCSZ pcszSourceFile,
192	unsigned long ulLine,
193	PCSZ pcszFunction)
194	{
195	PSZ pszReturn = NULL;
196	ULONG ulLength = 0;
197
198	if ( (pcszSource)
199	&& (ulLength = strlen(pcszSource))
200	)
201	{
202	if (pszReturn = (PSZ)memdMalloc(ulLength + 1,
203	pcszSourceFile, // fixed V0.9.16 (2001-12-08) [umoeller]
204	ulLine,
205	pcszFunction))
206	memcpy(pszReturn, pcszSource, ulLength + 1);
207	}
208
209	if (pulLength)
210	*pulLength = ulLength;
211
212	return (pszReturn);
213	}
214
215	#endif // __DEBUG_MALLOC_ENABLED__
216
217	/*
218	*@@ strhdup:
219	* like strdup, but this one doesn't crash if pszSource
220	* is NULL, but returns NULL also. In addition, this
221	* can report the length of the string (V0.9.16).
222	*
223	*@@added V0.9.0 [umoeller]
224	*@@changed V0.9.16 (2001-10-25) [umoeller]: added pulLength
225	*/
226
227	PSZ strhdup(PCSZ pcszSource,
228	unsigned long *pulLength) // out: length of string excl. null terminator (ptr can be NULL)
229	{
230	PSZ pszReturn = NULL;
231	ULONG ulLength = 0;
232
233	if ( (pcszSource)
234	&& (ulLength = strlen(pcszSource))
235	)
236	{
237	if (pszReturn = (PSZ)malloc(ulLength + 1))
238	memcpy(pszReturn, pcszSource, ulLength + 1);
239	}
240
241	if (pulLength)
242	*pulLength = ulLength;
243
244	return (pszReturn);
245	}
246
247	/*
248	*@@ strhcmp:
249	* better strcmp. This doesn't crash if any of the
250	* string pointers are NULL, but returns a proper
251	* value then.
252	*
253	* Besides, this is guaranteed to only return -1, 0,
254	* or +1, while strcmp can return any positive or
255	* negative value. This is useful for tree comparison
256	* funcs.
257	*
258	*@@added V0.9.9 (2001-02-16) [umoeller]
259	*/
260
261	int strhcmp(PCSZ p1, PCSZ p2)
262	{
263	if (p1 && p2)
264	{
265	int i = strcmp(p1, p2);
266	if (i < 0) return (-1);
267	if (i > 0) return (+1);
268	}
269	else if (p1)
270	// but p2 is NULL: p1 greater than p2 then
271	return (+1);
272	else if (p2)
273	// but p1 is NULL: p1 less than p2 then
274	return (-1);
275
276	// return 0 if strcmp returned 0 above or both strings are NULL
277	return (0);
278	}
279
280	/*
281	*@@ strhicmp:
282	* like strhcmp, but compares without respect
283	* to case.
284	*
285	*@@added V0.9.9 (2001-04-07) [umoeller]
286	*/
287
288	int strhicmp(PCSZ p1, PCSZ p2)
289	{
290	if (p1 && p2)
291	{
292	int i = stricmp(p1, p2);
293	if (i < 0) return (-1);
294	if (i > 0) return (+1);
295	}
296	else if (p1)
297	// but p2 is NULL: p1 greater than p2 then
298	return (+1);
299	else if (p2)
300	// but p1 is NULL: p1 less than p2 then
301	return (-1);
302
303	// return 0 if strcmp returned 0 above or both strings are NULL
304	return (0);
305	}
306
307	/*
308	*@@ strhistr:
309	* like strstr, but case-insensitive.
310	*
311	*@@changed V0.9.0 [umoeller]: crashed if null pointers were passed, thanks Rdiger Ihle
312	*/
313
314	PSZ strhistr(PCSZ string1, PCSZ string2)
315	{
316	PSZ prc = NULL;
317
318	if ((string1) && (string2))
319	{
320	PSZ pszSrchIn = strdup(string1);
321	PSZ pszSrchFor = strdup(string2);
322
323	if ((pszSrchIn) && (pszSrchFor))
324	{
325	strupr(pszSrchIn);
326	strupr(pszSrchFor);
327
328	prc = strstr(pszSrchIn, pszSrchFor);
329	if (prc)
330	{
331	// prc now has the first occurence of the string,
332	// but in pszSrchIn; we need to map this
333	// return value to the original string
334	prc = (prc-pszSrchIn) // offset in pszSrchIn
335	+ (PSZ)string1;
336	}
337	}
338	if (pszSrchFor)
339	free(pszSrchFor);
340	if (pszSrchIn)
341	free(pszSrchIn);
342	}
343	return (prc);
344	}
345
346	/*
347	*@@ strhncpy0:
348	* like strncpy, but always appends a 0 character.
349	*
350	*@@changed V0.9.16 (2002-01-09) [umoeller]: fixed crash on null pszSource
351	*/
352
353	ULONG strhncpy0(PSZ pszTarget,
354	PCSZ pszSource,
355	ULONG cbSource)
356	{
357	ULONG ul = 0;
358	PSZ pTarget = pszTarget,
359	pSource;
360
361	if (pSource = (PSZ)pszSource) // V0.9.16 (2002-01-09) [umoeller]
362	{
363	for (ul = 0; ul < cbSource; ul++)
364	if (*pSource)
365	pTarget++ = pSource++;
366	else
367	break;
368	}
369
370	*pTarget = 0;
371
372	return (ul);
373	}
374
375	/*
376	*@@ strhSize:
377	* returns the size of the given string, which
378	* is the memory required to allocate a copy,
379	* including the null terminator.
380	*
381	* Returns 0 if pcsz is NULL or points to a
382	* null character.
383	*
384	*@@added V0.9.18 (2002-02-13) [umoeller]
385	*/
386
387	ULONG strhSize(PCSZ pcsz)
388	{
389	if (pcsz && *pcsz)
390	return (strlen(pcsz) + 1);
391
392	return (0);
393	}
394
395	/*
396	* strhCount:
397	* this counts the occurences of c in pszSearch.
398	*/
399
400	ULONG strhCount(PCSZ pszSearch,
401	CHAR c)
402	{
403	PSZ p = (PSZ)pszSearch;
404	ULONG ulCount = 0;
405	while (TRUE)
406	{
407	p = strchr(p, c);
408	if (p)
409	{
410	ulCount++;
411	p++;
412	}
413	else
414	break;
415	}
416	return (ulCount);
417	}
418
419	/*
420	*@@ strhIsDecimal:
421	* returns TRUE if psz consists of decimal digits only.
422	*/
423
424	BOOL strhIsDecimal(PSZ psz)
425	{
426	PSZ p = psz;
427	while (*p != 0)
428	{
429	if (isdigit(*p) == 0)
430	return (FALSE);
431	p++;
432	}
433
434	return (TRUE);
435	}
436
437	#ifdef __DEBUG_MALLOC_ENABLED__
438
439	/*
440	*@@ strhSubstrDebug:
441	* memory debug version of strhSubstr.
442	*
443	*@@added V0.9.14 (2001-08-01) [umoeller]
444	*/
445
446	PSZ strhSubstrDebug(PCSZ pBegin, // in: first char
447	PCSZ pEnd, // in: last char (not included)
448	PCSZ pcszSourceFile,
449	unsigned long ulLine,
450	PCSZ pcszFunction)
451	{
452	PSZ pszSubstr = NULL;
453
454	if (pEnd > pBegin) // V0.9.9 (2001-04-04) [umoeller]
455	{
456	ULONG cbSubstr = (pEnd - pBegin);
457	if (pszSubstr = (PSZ)memdMalloc(cbSubstr + 1,
458	pcszSourceFile,
459	ulLine,
460	pcszFunction))
461	{
462	// strhncpy0(pszSubstr, pBegin, cbSubstr);
463	memcpy(pszSubstr, pBegin, cbSubstr); // V0.9.9 (2001-04-04) [umoeller]
464	*(pszSubstr + cbSubstr) = '\0';
465	}
466	}
467
468	return (pszSubstr);
469	}
470
471	#endif // __DEBUG_MALLOC_ENABLED__
472
473	/*
474	*@@ strhSubstr:
475	* this creates a new PSZ containing the string
476	* from pBegin to pEnd, excluding the pEnd character.
477	* The new string is null-terminated. The caller
478	* must free() the new string after use.
479	*
480	* Example:
481	+ "1234567890"
482	+ ^ ^
483	+ p1 p2
484	+ strhSubstr(p1, p2)
485	* would return a new string containing "2345678".
486	*
487	*@@changed V0.9.9 (2001-04-04) [umoeller]: fixed crashes with invalid pointers
488	*@@changed V0.9.9 (2001-04-04) [umoeller]: now using memcpy for speed
489	*/
490
491	PSZ strhSubstr(PCSZ pBegin, // in: first char
492	PCSZ pEnd) // in: last char (not included)
493	{
494	PSZ pszSubstr = NULL;
495
496	if (pEnd > pBegin) // V0.9.9 (2001-04-04) [umoeller]
497	{
498	ULONG cbSubstr = (pEnd - pBegin);
499	if (pszSubstr = (PSZ)malloc(cbSubstr + 1))
500	{
501	memcpy(pszSubstr, pBegin, cbSubstr); // V0.9.9 (2001-04-04) [umoeller]
502	*(pszSubstr + cbSubstr) = '\0';
503	}
504	}
505
506	return (pszSubstr);
507	}
508
509	/*
510	*@@ strhExtract:
511	* searches pszBuf for the cOpen character and returns
512	* the data in between cOpen and cClose, excluding
513	* those two characters, in a newly allocated buffer
514	* which you must free() afterwards.
515	*
516	* Spaces and newlines/linefeeds are skipped.
517	*
518	* If the search was successful, the new buffer
519	* is returned and, if (ppEnd != NULL), *ppEnd points
520	* to the first character after the cClose character
521	* found in the buffer.
522	*
523	* If the search was not successful, NULL is
524	* returned, and *ppEnd is unchanged.
525	*
526	* If another cOpen character is found before
527	* cClose, matching cClose characters will be skipped.
528	* You can therefore nest the cOpen and cClose
529	* characters.
530	*
531	* This function ignores cOpen and cClose characters
532	* in C-style comments and strings surrounded by
533	* double quotes.
534	*
535	* Example:
536	+ PSZ pszBuf = "KEYWORD { --blah-- } next",
537	+ pEnd;
538	+ strhExtract(pszBuf,
539	+ '{', '}',
540	+ &pEnd)
541	* would return a new buffer containing " --blah-- ",
542	* and ppEnd would afterwards point to the space
543	* before "next" in the static buffer.
544	*
545	*@@added V0.9.0 [umoeller]
546	*/
547
548	PSZ strhExtract(PSZ pszBuf, // in: search buffer
549	CHAR cOpen, // in: opening char
550	CHAR cClose, // in: closing char
551	PSZ *ppEnd) // out: if != NULL, receives first character after closing char
552	{
553	PSZ pszReturn = NULL;
554
555	if (pszBuf)
556	{
557	PSZ pOpen;
558	if (pOpen = strchr(pszBuf, cOpen))
559	{
560	// opening char found:
561	// now go thru the whole rest of the buffer
562	PSZ p = pOpen+1;
563	LONG lLevel = 1; // if this goes 0, we're done
564	while (*p)
565	{
566	if (*p == cOpen)
567	lLevel++;
568	else if (*p == cClose)
569	{
570	lLevel--;
571	if (lLevel <= 0)
572	{
573	// matching closing bracket found:
574	// extract string
575	pszReturn = strhSubstr(pOpen+1, // after cOpen
576	p); // excluding cClose
577	if (ppEnd)
578	*ppEnd = p+1;
579	break; // while (*p)
580	}
581	}
582	else if (*p == '\"')
583	{
584	// beginning of string:
585	PSZ p2 = p+1;
586	// find end of string
587	while ((p2) && (p2 != '\"'))
588	p2++;
589
590	if (*p2 == '\"')
591	// closing quote found:
592	// search on after that
593	p = p2; // raised below
594	else
595	break; // while (*p)
596	}
597
598	p++;
599	}
600	}
601	}
602
603	return (pszReturn);
604	}
605
606	/*
607	*@@ strhQuote:
608	* similar to strhExtract, except that
609	* opening and closing chars are the same,
610	* and therefore no nesting is possible.
611	* Useful for extracting stuff between
612	* quotes.
613	*
614	*@@added V0.9.0 [umoeller]
615	*/
616
617	PSZ strhQuote(PSZ pszBuf,
618	CHAR cQuote,
619	PSZ *ppEnd)
620	{
621	PSZ pszReturn = NULL,
622	p1 = NULL;
623	if ((p1 = strchr(pszBuf, cQuote)))
624	{
625	PSZ p2;
626	if (p2 = strchr(p1+1, cQuote))
627	{
628	pszReturn = strhSubstr(p1+1, p2);
629	if (ppEnd)
630	// store closing char
631	*ppEnd = p2 + 1;
632	}
633	}
634
635	return (pszReturn);
636	}
637
638	/*
639	*@@ strhStrip:
640	* removes all double spaces.
641	* This copies within the "psz" buffer.
642	* If any double spaces are found, the
643	* string will be shorter than before,
644	* but the buffer is _not_ reallocated,
645	* so there will be unused bytes at the
646	* end.
647	*
648	* Returns the number of spaces removed.
649	*
650	*@@added V0.9.0 [umoeller]
651	*/
652
653	ULONG strhStrip(PSZ psz) // in/out: string
654	{
655	PSZ p;
656	ULONG cb = strlen(psz),
657	ulrc = 0;
658
659	for (p = psz; p < psz+cb; p++)
660	{
661	if ((p == ' ') && ((p+1) == ' '))
662	{
663	PSZ p2 = p;
664	while (*p2)
665	{
666	p2 = (p2+1);
667	p2++;
668	}
669	cb--;
670	p--;
671	ulrc++;
672	}
673	}
674	return (ulrc);
675	}
676
677	/*
678	*@@ strhins:
679	* this inserts one string into another.
680	*
681	* pszInsert is inserted into pszBuffer at offset
682	* ulInsertOfs (which counts from 0).
683	*
684	* A newly allocated string is returned. pszBuffer is
685	* not changed. The new string should be free()'d after
686	* use.
687	*
688	* Upon errors, NULL is returned.
689	*
690	*@@changed V0.9.0 [umoeller]: completely rewritten.
691	*/
692
693	PSZ strhins(PCSZ pcszBuffer,
694	ULONG ulInsertOfs,
695	PCSZ pcszInsert)
696	{
697	PSZ pszNew = NULL;
698
699	if ((pcszBuffer) && (pcszInsert))
700	{
701	do {
702	ULONG cbBuffer = strlen(pcszBuffer);
703	ULONG cbInsert = strlen(pcszInsert);
704
705	// check string length
706	if (ulInsertOfs > cbBuffer + 1)
707	break; // do
708
709	// OK, let's go.
710	pszNew = (PSZ)malloc(cbBuffer + cbInsert + 1); // additional null terminator
711
712	// copy stuff before pInsertPos
713	memcpy(pszNew,
714	pcszBuffer,
715	ulInsertOfs);
716	// copy string to be inserted
717	memcpy(pszNew + ulInsertOfs,
718	pcszInsert,
719	cbInsert);
720	// copy stuff after pInsertPos
721	strcpy(pszNew + ulInsertOfs + cbInsert,
722	pcszBuffer + ulInsertOfs);
723	} while (FALSE);
724	}
725
726	return (pszNew);
727	}
728
729	/*
730	*@@ strhFindReplace:
731	* wrapper around xstrFindReplace to work with C strings.
732	* Note that *ppszBuf can get reallocated and must
733	* be free()'able.
734	*
735	* Repetitive use of this wrapper is not recommended
736	* because it is considerably slower than xstrFindReplace.
737	*
738	*@@added V0.9.6 (2000-11-01) [umoeller]
739	*@@changed V0.9.7 (2001-01-15) [umoeller]: renamed from strhrpl
740	*/
741
742	ULONG strhFindReplace(PSZ *ppszBuf, // in/out: string
743	PULONG pulOfs, // in: where to begin search (0 = start);
744	// out: ofs of first char after replacement string
745	PCSZ pcszSearch, // in: search string; cannot be NULL
746	PCSZ pcszReplace) // in: replacement string; cannot be NULL
747	{
748	ULONG ulrc = 0;
749	XSTRING xstrBuf,
750	xstrFind,
751	xstrReplace;
752	size_t ShiftTable[256];
753	BOOL fRepeat = FALSE;
754	xstrInitSet(&xstrBuf, *ppszBuf);
755	// reallocated and returned, so we're safe
756	xstrInitSet(&xstrFind, (PSZ)pcszSearch);
757	xstrInitSet(&xstrReplace, (PSZ)pcszReplace);
758	// these two are never freed, so we're safe too
759
760	if ((ulrc = xstrFindReplace(&xstrBuf,
761	pulOfs,
762	&xstrFind,
763	&xstrReplace,
764	ShiftTable,
765	&fRepeat)))
766	// replaced:
767	*ppszBuf = xstrBuf.psz;
768
769	return (ulrc);
770	}
771
772	/*
773	* strhWords:
774	* returns the no. of words in "psz".
775	* A string is considered a "word" if
776	* it is surrounded by spaces only.
777	*
778	*@@added V0.9.0 [umoeller]
779	*/
780
781	ULONG strhWords(PSZ psz)
782	{
783	PSZ p;
784	ULONG cb = strlen(psz),
785	ulWords = 0;
786	if (cb > 1)
787	{
788	ulWords = 1;
789	for (p = psz; p < psz+cb; p++)
790	if (*p == ' ')
791	ulWords++;
792	}
793	return (ulWords);
794	}
795
796	/*
797	*@@ strhGetWord:
798	* finds word boundaries.
799	*
800	* *ppszStart is used as the beginning of the
801	* search.
802	*
803	* If a word is found, *ppszStart is set to
804	* the first character of the word which was
805	* found and *ppszEnd receives the address
806	* of the first character _after_ the word,
807	* which is probably a space or a \n or \r char.
808	* We then return TRUE.
809	*
810	* The search is stopped if a null character
811	* is found or pLimit is reached. In that case,
812	* FALSE is returned.
813	*
814	*@@added V0.9.1 (2000-02-13) [umoeller]
815	*/
816
817	BOOL strhGetWord(PSZ *ppszStart, // in: start of search,
818	// out: start of word (if TRUE is returned)
819	PCSZ pLimit, // in: ptr to last char after *ppszStart to be
820	// searched; if the word does not end before
821	// or with this char, FALSE is returned
822	PCSZ pcszBeginChars, // stringh.h defines STRH_BEGIN_CHARS
823	PCSZ pcszEndChars, // stringh.h defines STRH_END_CHARS
824	PSZ *ppszEnd) // out: first char _after_ word
825	// (if TRUE is returned)
826	{
827	// characters after which a word can be started
828	// PCSZ pcszBeginChars = "\x0d\x0a ";
829	// PCSZ pcszEndChars = "\x0d\x0a /-";
830
831	PSZ pStart = *ppszStart;
832
833	// find start of word
834	while ( (pStart < (PSZ)pLimit)
835	&& (strchr(pcszBeginChars, *pStart))
836	)
837	// if char is a "before word" char: go for next
838	pStart++;
839
840	if (pStart < (PSZ)pLimit)
841	{
842	// found a valid "word start" character
843	// (which is not in pcszBeginChars):
844
845	// find end of word
846	PSZ pEndOfWord = pStart;
847	while ( (pEndOfWord <= (PSZ)pLimit)
848	&& (strchr(pcszEndChars, *pEndOfWord) == 0)
849	)
850	// if char is not an "end word" char: go for next
851	pEndOfWord++;
852
853	if (pEndOfWord <= (PSZ)pLimit)
854	{
855	// whoa, got a word:
856	*ppszStart = pStart;
857	*ppszEnd = pEndOfWord;
858	return (TRUE);
859	}
860	}
861
862	return (FALSE);
863	}
864
865	/*
866	*@@ strhIsWord:
867	* returns TRUE if p points to a "word"
868	* in pcszBuf.
869	*
870	* p is considered a word if the character _before_
871	* it is in pcszBeginChars and the char _after_
872	* it (i.e. *(p+cbSearch)) is in pcszEndChars.
873	*
874	*@@added V0.9.6 (2000-11-12) [umoeller]
875	*@@changed V0.9.18 (2002-02-23) [umoeller]: fixed end char check
876	*/
877
878	BOOL strhIsWord(PCSZ pcszBuf,
879	PCSZ p, // in: start of word
880	ULONG cbSearch, // in: length of word
881	PCSZ pcszBeginChars, // suggestion: "\x0d\x0a ()/\\-,."
882	PCSZ pcszEndChars) // suggestion: "\x0d\x0a ()/\\-,.:;"
883	{
884	// check previous char
885	if ( (p == pcszBuf)
886	\|\| (strchr(pcszBeginChars, *(p-1)))
887	)
888	{
889	// OK, valid begin char:
890	// check end char
891	CHAR cNextChar;
892	if (!(cNextChar = p[cbSearch]))
893	// null terminator:
894	return TRUE;
895	else
896	{
897	// not null terminator: check if char is
898	// in the list of valid end chars
899	if (strchr(pcszEndChars, cNextChar))
900	{
901	// OK, is end char: avoid doubles of that char,
902	// but allow spaces
903	// fixed V0.9.18 (2002-02-23) [umoeller]
904	CHAR cNextNext = p[cbSearch + 1];
905	if ( (cNextNext != cNextChar)
906	\|\| (cNextNext == ' ')
907	\|\| (cNextNext == 0)
908	)
909	return TRUE;
910	}
911	}
912	}
913
914	return FALSE;
915	}
916
917	/*
918	*@@ strhFindWord:
919	* searches for pszSearch in pszBuf, which is
920	* returned if found (or NULL if not).
921	*
922	* As opposed to strstr, this finds pszSearch
923	* only if it is a "word". A search string is
924	* considered a word if the character _before_
925	* it is in pcszBeginChars and the char _after_
926	* it is in pcszEndChars.
927	*
928	* Example:
929	+ strhFindWord("This is an example.", "is");
930	+ returns ...........^ this, but not the "is" in "This".
931	*
932	* The algorithm here uses strstr to find pszSearch in pszBuf
933	* and performs additional "is-word" checks for each item found
934	* (by calling strhIsWord).
935	*
936	* Note that this function is fairly slow compared to xstrFindWord.
937	*
938	*@@added V0.9.0 (99-11-08) [umoeller]
939	*@@changed V0.9.0 (99-11-10) [umoeller]: tried second algorithm, reverted to original...
940	*/
941
942	PSZ strhFindWord(PCSZ pszBuf,
943	PCSZ pszSearch,
944	PCSZ pcszBeginChars, // suggestion: "\x0d\x0a ()/\\-,."
945	PCSZ pcszEndChars) // suggestion: "\x0d\x0a ()/\\-,.:;"
946	{
947	PSZ pszReturn = 0;
948	ULONG cbBuf = strlen(pszBuf),
949	cbSearch = strlen(pszSearch);
950
951	if ((cbBuf) && (cbSearch))
952	{
953	PCSZ p = pszBuf;
954
955	do // while p
956	{
957	p = strstr(p, pszSearch);
958	if (p)
959	{
960	// string found:
961	// check if that's a word
962
963	if (strhIsWord(pszBuf,
964	p,
965	cbSearch,
966	pcszBeginChars,
967	pcszEndChars))
968	{
969	// valid end char:
970	pszReturn = (PSZ)p;
971	break;
972	}
973
974	p += cbSearch;
975	}
976	} while (p);
977
978	}
979	return (pszReturn);
980	}
981
982	/*
983	*@@ strhFindEOL:
984	* returns a pointer to the next \r, \n or null character
985	* following pszSearchIn. Stores the offset in *pulOffset.
986	*
987	* This should never return NULL because at some point,
988	* there will be a null byte in your string.
989	*
990	*@@added V0.9.4 (2000-07-01) [umoeller]
991	*/
992
993	PSZ strhFindEOL(PCSZ pcszSearchIn, // in: where to search
994	PULONG pulOffset) // out: offset (ptr can be NULL)
995	{
996	PCSZ p = pcszSearchIn,
997	prc = 0;
998	while (TRUE)
999	{
1000	if ( (p == '\r') \|\| (p == '\n') \|\| (*p == 0) )
1001	{
1002	prc = p;
1003	break;
1004	}
1005	p++;
1006	}
1007
1008	if ((pulOffset) && (prc))
1009	*pulOffset = prc - pcszSearchIn;
1010
1011	return ((PSZ)prc);
1012	}
1013
1014	/*
1015	*@@ strhFindNextLine:
1016	* like strhFindEOL, but this returns the character
1017	* _after_ \r or \n. Note that this might return
1018	* a pointer to terminating NULL character also.
1019	*/
1020
1021	PSZ strhFindNextLine(PSZ pszSearchIn, PULONG pulOffset)
1022	{
1023	PSZ pEOL = strhFindEOL(pszSearchIn, NULL);
1024	// pEOL now points to the \r char or the terminating 0 byte;
1025	// if not null byte, advance pointer
1026	PSZ pNextLine = pEOL;
1027	if (*pNextLine == '\r')
1028	pNextLine++;
1029	if (*pNextLine == '\n')
1030	pNextLine++;
1031	if (pulOffset)
1032	*pulOffset = pNextLine - pszSearchIn;
1033	return (pNextLine);
1034	}
1035
1036	/*
1037	*@@ strhBeautifyTitle:
1038	* replaces all line breaks (0xd, 0xa) with spaces.
1039	*
1040	*@@changed V0.9.12 (2001-05-17) [pr]: multiple line break chars. end up as only 1 space
1041	*/
1042
1043	BOOL strhBeautifyTitle(PSZ psz)
1044	{
1045	BOOL rc = FALSE;
1046	CHAR *p = psz;
1047
1048	while(*p)
1049	if ( (*p == '\r')
1050	\|\| (*p == '\n')
1051	)
1052	{
1053	rc = TRUE;
1054	if ( (p != psz)
1055	&& (p[-1] == ' ')
1056	)
1057	memmove(p, p + 1, strlen(p));
1058	else
1059	*p++ = ' ';
1060	}
1061	else
1062	p++;
1063
1064	return (rc);
1065	}
1066
1067	/*
1068	* strhFindAttribValue:
1069	* searches for pszAttrib in pszSearchIn; if found,
1070	* returns the first character after the "=" char.
1071	* If "=" is not found, a space, \r, and \n are
1072	* also accepted. This function searches without
1073	* respecting case.
1074	*
1075	* <B>Example:</B>
1076	+ strhFindAttribValue("<PAGE BLAH=\"data\">", "BLAH")
1077	+
1078	+ returns ....................... ^ this address.
1079	*
1080	*@@added V0.9.0 [umoeller]
1081	*@@changed V0.9.3 (2000-05-19) [umoeller]: some speed optimizations
1082	*@@changed V0.9.12 (2001-05-22) [umoeller]: fixed space bug, thanks Yuri Dario
1083	*/
1084
1085	PSZ strhFindAttribValue(const char pszSearchIn, const char pszAttrib)
1086	{
1087	PSZ prc = 0;
1088	PSZ pszSearchIn2, p;
1089	ULONG cbAttrib = strlen(pszAttrib),
1090	ulLength = strlen(pszSearchIn);
1091
1092	// use alloca(), so memory is freed on function exit
1093	pszSearchIn2 = (PSZ)alloca(ulLength + 1);
1094	memcpy(pszSearchIn2, pszSearchIn, ulLength + 1);
1095
1096	// 1) find token, (space char, \n, \r, \t)
1097	p = strtok(pszSearchIn2, " \n\r\t");
1098	while (p)
1099	{
1100	CHAR c2;
1101	PSZ pOrig;
1102
1103	// check tag name
1104	if (!strnicmp(p, pszAttrib, cbAttrib))
1105	{
1106	// position in original string
1107	pOrig = (PSZ)pszSearchIn + (p - pszSearchIn2);
1108
1109	// yes:
1110	prc = pOrig + cbAttrib;
1111	c2 = *prc;
1112	while ( ( (c2 == ' ')
1113	\|\| (c2 == '=')
1114	\|\| (c2 == '\n')
1115	\|\| (c2 == '\r')
1116	)
1117	&& (c2 != 0)
1118	)
1119	c2 = *++prc;
1120
1121	break;
1122	}
1123
1124	p = strtok(NULL, " \n\r\t");
1125	}
1126
1127	return (prc);
1128	}
1129
1130	/* PSZ strhFindAttribValue(const char pszSearchIn, const char pszAttrib)
1131	{
1132	PSZ prc = 0;
1133	PSZ pszSearchIn2 = (PSZ)pszSearchIn,
1134	p,
1135	p2;
1136	ULONG cbAttrib = strlen(pszAttrib);
1137
1138	// 1) find space char
1139	while ((p = strchr(pszSearchIn2, ' ')))
1140	{
1141	CHAR c;
1142	p++;
1143	if (strlen(p) >= cbAttrib) // V0.9.9 (2001-03-27) [umoeller]
1144	{
1145	c = *(p+cbAttrib); // V0.9.3 (2000-05-19) [umoeller]
1146	// now check whether the p+strlen(pszAttrib)
1147	// is a valid end-of-tag character
1148	if ( (memicmp(p, (PVOID)pszAttrib, cbAttrib) == 0)
1149	&& ( (c == ' ')
1150	\|\| (c == '>')
1151	\|\| (c == '=')
1152	\|\| (c == '\r')
1153	\|\| (c == '\n')
1154	\|\| (c == 0)
1155	)
1156	)
1157	{
1158	// yes:
1159	CHAR c2;
1160	p2 = p + cbAttrib;
1161	c2 = *p2;
1162	while ( ( (c2 == ' ')
1163	\|\| (c2 == '=')
1164	\|\| (c2 == '\n')
1165	\|\| (c2 == '\r')
1166	)
1167	&& (c2 != 0)
1168	)
1169	c2 = *++p2;
1170
1171	prc = p2;
1172	break; // first while
1173	}
1174	}
1175	else
1176	break;
1177
1178	pszSearchIn2++;
1179	}
1180	return (prc);
1181	} */
1182
1183	/*
1184	* strhGetNumAttribValue:
1185	* stores the numerical parameter value of an HTML-style
1186	* tag in *pl.
1187	*
1188	* Returns the address of the tag parameter in the
1189	* search buffer, if found, or NULL.
1190	*
1191	* <B>Example:</B>
1192	+ strhGetNumAttribValue("<PAGE BLAH=123>, "BLAH", &l);
1193	*
1194	* stores 123 in the "l" variable.
1195	*
1196	*@@added V0.9.0 [umoeller]
1197	*@@changed V0.9.9 (2001-04-04) [umoeller]: this failed on "123" strings in quotes, fixed
1198	*/
1199
1200	PSZ strhGetNumAttribValue(const char *pszSearchIn, // in: where to search
1201	const char *pszTag, // e.g. "INDEX"
1202	PLONG pl) // out: numerical value
1203	{
1204	PSZ pParam;
1205	if ((pParam = strhFindAttribValue(pszSearchIn, pszTag)))
1206	{
1207	if ( (*pParam == '\"')
1208	\|\| (*pParam == '\'')
1209	)
1210	pParam++; // V0.9.9 (2001-04-04) [umoeller]
1211
1212	sscanf(pParam, "%ld", pl);
1213	}
1214
1215	return (pParam);
1216	}
1217
1218	/*
1219	* strhGetTextAttr:
1220	* retrieves the attribute value of a textual HTML-style tag
1221	* in a newly allocated buffer, which is returned,
1222	* or NULL if attribute not found.
1223	* If an attribute value is to contain spaces, it
1224	* must be enclosed in quotes.
1225	*
1226	* The offset of the attribute data in pszSearchIn is
1227	* returned in *pulOffset so that you can do multiple
1228	* searches.
1229	*
1230	* This returns a new buffer, which should be free()'d after use.
1231	*
1232	* <B>Example:</B>
1233	+ ULONG ulOfs = 0;
1234	+ strhGetTextAttr("<PAGE BLAH="blublub">, "BLAH", &ulOfs)
1235	+ ............^ ulOfs
1236	*
1237	* returns a new string with the value "blublub" (without
1238	* quotes) and sets ulOfs to 12.
1239	*
1240	*@@added V0.9.0 [umoeller]
1241	*/
1242
1243	PSZ strhGetTextAttr(const char *pszSearchIn,
1244	const char *pszTag,
1245	PULONG pulOffset) // out: offset where found
1246	{
1247	PSZ pParam,
1248	pParam2,
1249	prc = NULL;
1250	ULONG ulCount = 0;
1251	LONG lNestingLevel = 0;
1252
1253	if ((pParam = strhFindAttribValue(pszSearchIn, pszTag)))
1254	{
1255	// determine end character to search for: a space
1256	CHAR cEnd = ' ';
1257	if (*pParam == '\"')
1258	{
1259	// or, if the data is enclosed in quotes, a quote
1260	cEnd = '\"';
1261	pParam++;
1262	}
1263
1264	if (pulOffset)
1265	// store the offset
1266	(*pulOffset) = pParam - (PSZ)pszSearchIn;
1267
1268	// now find end of attribute
1269	pParam2 = pParam;
1270	while (*pParam)
1271	{
1272	if (*pParam == cEnd)
1273	// end character found
1274	break;
1275	else if (*pParam == '<')
1276	// yet another opening tag found:
1277	// this is probably some "<" in the attributes
1278	lNestingLevel++;
1279	else if (*pParam == '>')
1280	{
1281	lNestingLevel--;
1282	if (lNestingLevel < 0)
1283	// end of tag found:
1284	break;
1285	}
1286	ulCount++;
1287	pParam++;
1288	}
1289
1290	// copy attribute to new buffer
1291	if (ulCount)
1292	{
1293	prc = (PSZ)malloc(ulCount+1);
1294	memcpy(prc, pParam2, ulCount);
1295	*(prc+ulCount) = 0;
1296	}
1297	}
1298	return (prc);
1299	}
1300
1301	/*
1302	* strhFindEndOfTag:
1303	* returns a pointer to the ">" char
1304	* which seems to terminate the tag beginning
1305	* after pszBeginOfTag.
1306	*
1307	* If additional "<" chars are found, we look
1308	* for additional ">" characters too.
1309	*
1310	* Note: You must pass the address of the opening
1311	* '<' character to this function.
1312	*
1313	* Example:
1314	+ PSZ pszTest = "<BODY ATTR=\"<BODY>\">";
1315	+ strhFindEndOfTag(pszTest)
1316	+ returns.................................^ this.
1317	*
1318	*@@added V0.9.0 [umoeller]
1319	*/
1320
1321	PSZ strhFindEndOfTag(const char *pszBeginOfTag)
1322	{
1323	PSZ p = (PSZ)pszBeginOfTag,
1324	prc = NULL;
1325	LONG lNestingLevel = 0;
1326
1327	while (*p)
1328	{
1329	if (*p == '<')
1330	// another opening tag found:
1331	lNestingLevel++;
1332	else if (*p == '>')
1333	{
1334	// closing tag found:
1335	lNestingLevel--;
1336	if (lNestingLevel < 1)
1337	{
1338	// corresponding: return this
1339	prc = p;
1340	break;
1341	}
1342	}
1343	p++;
1344	}
1345
1346	return (prc);
1347	}
1348
1349	/*
1350	* strhGetBlock:
1351	* this complex function searches the given string
1352	* for a pair of opening/closing HTML-style tags.
1353	*
1354	* If found, this routine returns TRUE and does
1355	* the following:
1356	*
1357	* 1) allocate a new buffer, copy the text
1358	* enclosed by the opening/closing tags
1359	* into it and set *ppszBlock to that
1360	* buffer;
1361	*
1362	* 2) if the opening tag has any attributes,
1363	* allocate another buffer, copy the
1364	* attributes into it and set *ppszAttrs
1365	* to that buffer; if no attributes are
1366	* found, *ppszAttrs will be NULL;
1367	*
1368	* 3) set *pulOffset to the offset from the
1369	* beginning of *ppszSearchIn where the
1370	* opening tag was found;
1371	*
1372	* 4) advance *ppszSearchIn to after the
1373	* closing tag, so that you can do
1374	* multiple searches without finding the
1375	* same tags twice.
1376	*
1377	* All buffers should be freed using free().
1378	*
1379	* This returns the following:
1380	* -- 0: no error
1381	* -- 1: tag not found at all (doesn't have to be an error)
1382	* -- 2: begin tag found, but no corresponding end tag found. This
1383	* is a real error.
1384	* -- 3: begin tag is not terminated by ">" (e.g. "<BEGINTAG whatever")
1385	*
1386	* <B>Example:</B>
1387	+ PSZ pSearch = "<PAGE INDEX=1>This is page 1.</PAGE>More text."
1388	+ PSZ pszBlock, pszAttrs;
1389	+ ULONG ulOfs;
1390	+ strhGetBlock(&pSearch, "PAGE", &pszBlock, &pszAttrs, &ulOfs)
1391	*
1392	* would do the following:
1393	*
1394	* 1) set pszBlock to a new string containing "This is page 1."
1395	* without quotes;
1396	*
1397	* 2) set pszAttrs to a new string containing "<PAGE INDEX=1>";
1398	*
1399	* 3) set ulOfs to 0, because "<PAGE" was found at the beginning;
1400	*
1401	* 4) pSearch would be advanced to point to the "More text"
1402	* string in the original buffer.
1403	*
1404	* Hey-hey. A one-shot function, fairly complicated, but indispensable
1405	* for HTML parsing.
1406	*
1407	*@@added V0.9.0 [umoeller]
1408	*@@changed V0.9.1 (2000-01-03) [umoeller]: fixed heap overwrites (thanks to string debugging)
1409	*@@changed V0.9.1 (2000-01-06) [umoeller]: changed prototype
1410	*@@changed V0.9.3 (2000-05-06) [umoeller]: NULL string check was missing
1411	*/
1412
1413	ULONG strhGetBlock(const char *pszSearchIn, // in: buffer to search
1414	PULONG pulSearchOffset, // in/out: offset where to start search (0 for beginning)
1415	PSZ pszTag,
1416	PSZ *ppszBlock, // out: block enclosed by the tags
1417	PSZ *ppszAttribs, // out: attributes of the opening tag
1418	PULONG pulOfsBeginTag, // out: offset from pszSearchIn where opening tag was found
1419	PULONG pulOfsBeginBlock) // out: offset from pszSearchIn where beginning of block was found
1420	{
1421	ULONG ulrc = 1;
1422	PSZ pszBeginTag = (PSZ)pszSearchIn + *pulSearchOffset,
1423	pszSearch2 = pszBeginTag,
1424	pszClosingTag;
1425	ULONG cbTag = strlen(pszTag);
1426
1427	// go thru the block and check all tags if it's the
1428	// begin tag we're looking for
1429	while ((pszBeginTag = strchr(pszBeginTag, '<')))
1430	{
1431	if (memicmp(pszBeginTag+1, pszTag, strlen(pszTag)) == 0)
1432	// yes: stop
1433	break;
1434	else
1435	pszBeginTag++;
1436	}
1437
1438	if (pszBeginTag)
1439	{
1440	// we found <TAG>:
1441	ULONG ulNestingLevel = 0;
1442
1443	PSZ pszEndOfBeginTag = strhFindEndOfTag(pszBeginTag);
1444	// strchr(pszBeginTag, '>');
1445	if (pszEndOfBeginTag)
1446	{
1447	// does the caller want the attributes?
1448	if (ppszAttribs)
1449	{
1450	// yes: then copy them
1451	ULONG ulAttrLen = pszEndOfBeginTag - pszBeginTag;
1452	PSZ pszAttrs = (PSZ)malloc(ulAttrLen + 1);
1453	strncpy(pszAttrs, pszBeginTag, ulAttrLen);
1454	// add terminating 0
1455	*(pszAttrs + ulAttrLen) = 0;
1456
1457	*ppszAttribs = pszAttrs;
1458	}
1459
1460	// output offset of where we found the begin tag
1461	if (pulOfsBeginTag)
1462	*pulOfsBeginTag = pszBeginTag - (PSZ)pszSearchIn;
1463
1464	// now find corresponding closing tag (e.g. "</BODY>"
1465	pszBeginTag = pszEndOfBeginTag+1;
1466	// now we're behind the '>' char of the opening tag
1467	// increase offset of that too
1468	if (pulOfsBeginBlock)
1469	*pulOfsBeginBlock = pszBeginTag - (PSZ)pszSearchIn;
1470
1471	// find next closing tag;
1472	// for the first run, pszSearch2 points to right
1473	// after the '>' char of the opening tag
1474	pszSearch2 = pszBeginTag;
1475	while ( (pszSearch2) // fixed V0.9.3 (2000-05-06) [umoeller]
1476	&& (pszClosingTag = strstr(pszSearch2, "<"))
1477	)
1478	{
1479	// if we have another opening tag before our closing
1480	// tag, we need to have several closing tags before
1481	// we're done
1482	if (memicmp(pszClosingTag+1, pszTag, cbTag) == 0)
1483	ulNestingLevel++;
1484	else
1485	{
1486	// is this ours?
1487	if ( (*(pszClosingTag+1) == '/')
1488	&& (memicmp(pszClosingTag+2, pszTag, cbTag) == 0)
1489	)
1490	{
1491	// we've found a matching closing tag; is
1492	// it ours?
1493	if (ulNestingLevel == 0)
1494	{
1495	// our closing tag found:
1496	// allocate mem for a new buffer
1497	// and extract all the text between
1498	// open and closing tags to it
1499	ULONG ulLen = pszClosingTag - pszBeginTag;
1500	if (ppszBlock)
1501	{
1502	PSZ pNew = (PSZ)malloc(ulLen + 1);
1503	strhncpy0(pNew, pszBeginTag, ulLen);
1504	*ppszBlock = pNew;
1505	}
1506
1507	// raise search offset to after the closing tag
1508	*pulSearchOffset = (pszClosingTag + cbTag + 1) - (PSZ)pszSearchIn;
1509
1510	ulrc = 0;
1511
1512	break;
1513	} else
1514	// not our closing tag:
1515	ulNestingLevel--;
1516	}
1517	}
1518	// no matching closing tag: search on after that
1519	pszSearch2 = strhFindEndOfTag(pszClosingTag);
1520	} // end while (pszClosingTag = strstr(pszSearch2, "<"))
1521
1522	if (!pszClosingTag)
1523	// no matching closing tag found:
1524	// return 2 (closing tag not found)
1525	ulrc = 2;
1526	} // end if (pszBeginTag)
1527	else
1528	// no matching ">" for opening tag found:
1529	ulrc = 3;
1530	}
1531
1532	return (ulrc);
1533	}
1534
1535	/* ******************************************************************
1536	*
1537	* Miscellaneous
1538	*
1539	********************************************************************/
1540
1541	/*
1542	*@@ strhArrayAppend:
1543	* this appends a string to a "string array".
1544	*
1545	* A string array is considered a sequence of
1546	* zero-terminated strings in memory. That is,
1547	* after each string's null-byte, the next
1548	* string comes up.
1549	*
1550	* This is useful for composing a single block
1551	* of memory from, say, list box entries, which
1552	* can then be written to OS2.INI in one flush.
1553	*
1554	* To append strings to such an array, call this
1555	* function for each string you wish to append.
1556	* This will re-allocate *ppszRoot with each call,
1557	* and update *pcbRoot, which then contains the
1558	* total size of all strings (including all null
1559	* terminators).
1560	*
1561	* Pass *pcbRoot to PrfSaveProfileData to have the
1562	* block saved.
1563	*
1564	* Note: On the first call, ppszRoot and pcbRoot
1565	* _must_ be both NULL, or this crashes.
1566	*
1567	*@@changed V0.9.13 (2001-06-21) [umoeller]: added cbNew
1568	*/
1569
1570	VOID strhArrayAppend(PSZ *ppszRoot, // in: root of array
1571	const char *pcszNew, // in: string to append
1572	ULONG cbNew, // in: size of that string or 0 to run strlen() here
1573	PULONG pcbRoot) // in/out: size of array
1574	{
1575	PSZ pszTemp;
1576
1577	if (!cbNew) // V0.9.13 (2001-06-21) [umoeller]
1578	cbNew = strlen(pcszNew);
1579
1580	pszTemp = (PSZ)malloc(*pcbRoot
1581	+ cbNew
1582	+ 1); // two null bytes
1583	if (*ppszRoot)
1584	{
1585	// not first loop: copy old stuff
1586	memcpy(pszTemp,
1587	*ppszRoot,
1588	*pcbRoot);
1589	free(*ppszRoot);
1590	}
1591	// append new string
1592	strcpy(pszTemp + *pcbRoot,
1593	pcszNew);
1594	// update root
1595	*ppszRoot = pszTemp;
1596	// update length
1597	*pcbRoot += cbNew + 1;
1598	}
1599
1600	/*
1601	*@@ strhCreateDump:
1602	* this dumps a memory block into a string
1603	* and returns that string in a new buffer.
1604	*
1605	* You must free() the returned PSZ after use.
1606	*
1607	* The output looks like the following:
1608	*
1609	+ 0000: FE FF 0E 02 90 00 00 00 ........
1610	+ 0008: FD 01 00 00 57 50 46 6F ....WPFo
1611	+ 0010: 6C 64 65 72 00 78 01 34 lder.x.4
1612	*
1613	* Each line is terminated with a newline (\n)
1614	* character only.
1615	*
1616	*@@added V0.9.1 (2000-01-22) [umoeller]
1617	*/
1618
1619	PSZ strhCreateDump(PBYTE pb, // in: start address of buffer
1620	ULONG ulSize, // in: size of buffer
1621	ULONG ulIndent) // in: indentation of every line
1622	{
1623	PSZ pszReturn = 0;
1624	XSTRING strReturn;
1625	CHAR szTemp[1000];
1626
1627	PBYTE pbCurrent = pb; // current byte
1628	ULONG ulCount = 0,
1629	ulCharsInLine = 0; // if this grows > 7, a new line is started
1630	CHAR szLine[400] = "",
1631	szAscii[30] = " "; // ASCII representation; filled for every line
1632	PSZ pszLine = szLine,
1633	pszAscii = szAscii;
1634
1635	xstrInit(&strReturn, (ulSize * 30) + ulIndent);
1636
1637	for (pbCurrent = pb;
1638	ulCount < ulSize;
1639	pbCurrent++, ulCount++)
1640	{
1641	if (ulCharsInLine == 0)
1642	{
1643	memset(szLine, ' ', ulIndent);
1644	pszLine += ulIndent;
1645	}
1646	pszLine += sprintf(pszLine, "%02lX ", (ULONG)*pbCurrent);
1647
1648	if ( (pbCurrent > 31) && (pbCurrent < 127) )
1649	// printable character:
1650	pszAscii = pbCurrent;
1651	else
1652	*pszAscii = '.';
1653	pszAscii++;
1654
1655	ulCharsInLine++;
1656	if ( (ulCharsInLine > 7) // 8 bytes added?
1657	\|\| (ulCount == ulSize-1) // end of buffer reached?
1658	)
1659	{
1660	// if we haven't had eight bytes yet,
1661	// fill buffer up to eight bytes with spaces
1662	ULONG ul2;
1663	for (ul2 = ulCharsInLine;
1664	ul2 < 8;
1665	ul2++)
1666	pszLine += sprintf(pszLine, " ");
1667
1668	sprintf(szTemp, "%04lX: %s %s\n",
1669	(ulCount & 0xFFFFFFF8), // offset in hex
1670	szLine, // bytes string
1671	szAscii); // ASCII string
1672	xstrcat(&strReturn, szTemp, 0);
1673
1674	// restart line buffer
1675	pszLine = szLine;
1676
1677	// clear ASCII buffer
1678	strcpy(szAscii, " ");
1679	pszAscii = szAscii;
1680
1681	// reset line counter
1682	ulCharsInLine = 0;
1683	}
1684	}
1685
1686	if (strReturn.cbAllocated)
1687	pszReturn = strReturn.psz;
1688
1689	return (pszReturn);
1690	}
1691
1692	/* ******************************************************************
1693	*
1694	* Fast string searches
1695	*
1696	********************************************************************/
1697
1698	#define ASSERT(a)
1699
1700	/*
1701	* The following code has been taken from the "Standard
1702	* Function Library", file sflfind.c, and only slightly
1703	* modified to conform to the rest of this file.
1704	*
1705	* Written: 96/04/24 iMatix SFL project team <sfl@imatix.com>
1706	* Revised: 98/05/04
1707	*
1708	* Copyright: Copyright (c) 1991-99 iMatix Corporation.
1709	*
1710	* The SFL Licence allows incorporating SFL code into other
1711	* programs, as long as the copyright is reprinted and the
1712	* code is marked as modified, so this is what we do.
1713	*/
1714
1715	/*
1716	*@@ strhmemfind:
1717	* searches for a pattern in a block of memory using the
1718	* Boyer-Moore-Horspool-Sunday algorithm.
1719	*
1720	* The block and pattern may contain any values; you must
1721	* explicitly provide their lengths. If you search for strings,
1722	* use strlen() on the buffers.
1723	*
1724	* Returns a pointer to the pattern if found within the block,
1725	* or NULL if the pattern was not found.
1726	*
1727	* This algorithm needs a "shift table" to cache data for the
1728	* search pattern. This table can be reused when performing
1729	* several searches with the same pattern.
1730	*
1731	* "shift" must point to an array big enough to hold 256 (8**2)
1732	* "size_t" values.
1733	*
1734	* If (*repeat_find == FALSE), the shift table is initialized.
1735	* So on the first search with a given pattern, *repeat_find
1736	* should be FALSE. This function sets it to TRUE after the
1737	* shift table is initialised, allowing the initialisation
1738	* phase to be skipped on subsequent searches.
1739	*
1740	* This function is most effective when repeated searches are
1741	* made for the same pattern in one or more large buffers.
1742	*
1743	* Example:
1744	*
1745	+ PSZ pszHaystack = "This is a sample string.",
1746	+ pszNeedle = "string";
1747	+ size_t shift[256];
1748	+ BOOL fRepeat = FALSE;
1749	+
1750	+ PSZ pFound = strhmemfind(pszHaystack,
1751	+ strlen(pszHaystack), // block size
1752	+ pszNeedle,
1753	+ strlen(pszNeedle), // pattern size
1754	+ shift,
1755	+ &fRepeat);
1756	*
1757	* Taken from the "Standard Function Library", file sflfind.c.
1758	* Copyright: Copyright (c) 1991-99 iMatix Corporation.
1759	* Slightly modified by umoeller.
1760	*
1761	*@@added V0.9.3 (2000-05-08) [umoeller]
1762	*/
1763
1764	void* strhmemfind(const void *in_block, // in: block containing data
1765	size_t block_size, // in: size of block in bytes
1766	const void *in_pattern, // in: pattern to search for
1767	size_t pattern_size, // in: size of pattern block
1768	size_t *shift, // in/out: shift table (search buffer)
1769	BOOL repeat_find) // in/out: if TRUE, shift is already initialized
1770	{
1771	size_t byte_nbr, // Distance through block
1772	match_size; // Size of matched part
1773	const unsigned char
1774	*match_base = NULL, // Base of match of pattern
1775	*match_ptr = NULL, // Point within current match
1776	*limit = NULL; // Last potiental match point
1777	const unsigned char
1778	block = (unsigned char ) in_block, // Concrete pointer to block data
1779	pattern = (unsigned char ) in_pattern; // Concrete pointer to search value
1780
1781	if ( (block == NULL)
1782	\|\| (pattern == NULL)
1783	\|\| (shift == NULL)
1784	)
1785	return (NULL);
1786
1787	// Pattern must be smaller or equal in size to string
1788	if (block_size < pattern_size)
1789	return (NULL); // Otherwise it's not found
1790
1791	if (pattern_size == 0) // Empty patterns match at start
1792	return ((void *)block);
1793
1794	// Build the shift table unless we're continuing a previous search
1795
1796	// The shift table determines how far to shift before trying to match
1797	// again, if a match at this point fails. If the byte after where the
1798	// end of our pattern falls is not in our pattern, then we start to
1799	// match again after that byte; otherwise we line up the last occurence
1800	// of that byte in our pattern under that byte, and try match again.
1801
1802	if (!repeat_find \|\| !*repeat_find)
1803	{
1804	for (byte_nbr = 0;
1805	byte_nbr < 256;
1806	byte_nbr++)
1807	shift[byte_nbr] = pattern_size + 1;
1808	for (byte_nbr = 0;
1809	byte_nbr < pattern_size;
1810	byte_nbr++)
1811	shift[(unsigned char)pattern[byte_nbr]] = pattern_size - byte_nbr;
1812
1813	if (repeat_find)
1814	*repeat_find = TRUE;
1815	}
1816
1817	// Search for the block, each time jumping up by the amount
1818	// computed in the shift table
1819
1820	limit = block + (block_size - pattern_size + 1);
1821	ASSERT (limit > block);
1822
1823	for (match_base = block;
1824	match_base < limit;
1825	match_base += shift[*(match_base + pattern_size)])
1826	{
1827	match_ptr = match_base;
1828	match_size = 0;
1829
1830	// Compare pattern until it all matches, or we find a difference
1831	while (*match_ptr++ == pattern[match_size++])
1832	{
1833	ASSERT (match_size <= pattern_size &&
1834	match_ptr == (match_base + match_size));
1835
1836	// If we found a match, return the start address
1837	if (match_size >= pattern_size)
1838	return ((void*)(match_base));
1839
1840	}
1841	}
1842	return (NULL); // Found nothing
1843	}
1844
1845	/*
1846	*@@ strhtxtfind:
1847	* searches for a case-insensitive text pattern in a string
1848	* using the Boyer-Moore-Horspool-Sunday algorithm. The string and
1849	* pattern are null-terminated strings. Returns a pointer to the pattern
1850	* if found within the string, or NULL if the pattern was not found.
1851	* Will match strings irrespective of case. To match exact strings, use
1852	* strhfind(). Will not work on multibyte characters.
1853	*
1854	* Examples:
1855	+ char *result;
1856	+
1857	+ result = strhtxtfind ("AbracaDabra", "cad");
1858	+ if (result)
1859	+ puts (result);
1860	+
1861	* Taken from the "Standard Function Library", file sflfind.c.
1862	* Copyright: Copyright (c) 1991-99 iMatix Corporation.
1863	* Slightly modified.
1864	*
1865	*@@added V0.9.3 (2000-05-08) [umoeller]
1866	*/
1867
1868	char* strhtxtfind (const char *string, // String containing data
1869	const char *pattern) // Pattern to search for
1870	{
1871	size_t
1872	shift [256]; // Shift distance for each value
1873	size_t
1874	string_size,
1875	pattern_size,
1876	byte_nbr, // Index into byte array
1877	match_size; // Size of matched part
1878	const char
1879	*match_base = NULL, // Base of match of pattern
1880	*match_ptr = NULL, // Point within current match
1881	*limit = NULL; // Last potiental match point
1882
1883	ASSERT (string); // Expect non-NULL pointers, but
1884	ASSERT (pattern); // fail gracefully if not debugging
1885	if (string == NULL \|\| pattern == NULL)
1886	return (NULL);
1887
1888	string_size = strlen (string);
1889	pattern_size = strlen (pattern);
1890
1891	// Pattern must be smaller or equal in size to string
1892	if (string_size < pattern_size)
1893	return (NULL); // Otherwise it cannot be found
1894
1895	if (pattern_size == 0) // Empty string matches at start
1896	return (char *) string;
1897
1898	// Build the shift table
1899
1900	// The shift table determines how far to shift before trying to match
1901	// again, if a match at this point fails. If the byte after where the
1902	// end of our pattern falls is not in our pattern, then we start to
1903	// match again after that byte; otherwise we line up the last occurence
1904	// of that byte in our pattern under that byte, and try match again.
1905
1906	for (byte_nbr = 0; byte_nbr < 256; byte_nbr++)
1907	shift [byte_nbr] = pattern_size + 1;
1908
1909	for (byte_nbr = 0; byte_nbr < pattern_size; byte_nbr++)
1910	shift [(unsigned char) tolower (pattern [byte_nbr])] = pattern_size - byte_nbr;
1911
1912	// Search for the string. If we don't find a match, move up by the
1913	// amount we computed in the shift table above, to find location of
1914	// the next potiental match.
1915
1916	limit = string + (string_size - pattern_size + 1);
1917	ASSERT (limit > string);
1918
1919	for (match_base = string;
1920	match_base < limit;
1921	match_base += shift [(unsigned char) tolower (*(match_base + pattern_size))])
1922	{
1923	match_ptr = match_base;
1924	match_size = 0;
1925
1926	// Compare pattern until it all matches, or we find a difference
1927	while (tolower (*match_ptr++) == tolower (pattern [match_size++]))
1928	{
1929	ASSERT (match_size <= pattern_size &&
1930	match_ptr == (match_base + match_size));
1931
1932	// If we found a match, return the start address
1933	if (match_size >= pattern_size)
1934	return ((char *)(match_base));
1935	}
1936	}
1937	return (NULL); // Found nothing
1938	}
1939

Note: See TracBrowser for help on using the repository browser.

Download in other formats: