X-Git-Url: http://www.privoxy.org/gitweb/?p=privoxy.git;a=blobdiff_plain;f=pcrs.c;h=ee51d0a39f1a5cdd4a4d37c03d8f948cc001c93e;hp=c2a4945e06844dcafbd7763a2f81b467f0e5ef5d;hb=41d37e938e75b5eaa032257fa1979f145d403b74;hpb=d2871ac82955cf0b901a6916da1f8ff66cd8fd72 diff --git a/pcrs.c b/pcrs.c index c2a4945e..ee51d0a3 100644 --- a/pcrs.c +++ b/pcrs.c @@ -1,4 +1,4 @@ -const char pcrs_rcs[] = "$Id: pcrs.c,v 1.13 2001/09/06 14:05:59 oes Exp $"; +const char pcrs_rcs[] = "$Id: pcrs.c,v 1.23 2006/12/29 17:53:05 fabiankeil Exp $"; /********************************************************************* * @@ -33,6 +33,54 @@ const char pcrs_rcs[] = "$Id: pcrs.c,v 1.13 2001/09/06 14:05:59 oes Exp $"; * * Revisions : * $Log: pcrs.c,v $ + * Revision 1.23 2006/12/29 17:53:05 fabiankeil + * Fixed gcc43 conversion warnings. + * + * Revision 1.22 2006/12/24 17:34:20 fabiankeil + * Add pcrs_strerror() message for PCRE_ERROR_MATCHLIMIT + * and give a hint why an error code might be unknown. + * + * Catch NULL subjects early in pcrs_execute(). + * + * Revision 1.21 2006/07/18 14:48:47 david__schmidt + * Reorganizing the repository: swapping out what was HEAD (the old 3.1 branch) + * with what was really the latest development (the v_3_0_branch branch) + * + * Revision 1.19.2.4 2005/05/07 21:50:55 david__schmidt + * A few memory leaks plugged (mostly on error paths) + * + * Revision 1.19.2.3 2003/12/04 12:32:45 oes + * Append a trailing nullbyte to result to facilitate string processing + * + * Revision 1.19.2.2 2002/10/08 16:22:28 oes + * Bugfix: Need to check validity of backreferences explicitly, + * because when max_matches are reached and matches is expanded, + * realloc() does not zero the memory. Fixes Bug # 606227 + * + * Revision 1.19.2.1 2002/08/10 11:23:40 oes + * Include prce.h via project.h, where the appropriate + * source will have been selected + * + * Revision 1.19 2002/03/08 14:47:48 oes + * Cosmetics + * + * Revision 1.18 2002/03/08 14:17:14 oes + * Fixing -Wconversion warnings + * + * Revision 1.17 2002/03/08 13:45:48 oes + * Hiding internal functions + * + * Revision 1.16 2001/11/30 21:32:14 jongfoster + * Fixing signed/unsigned comparison (Andreas please check this!) + * One tab->space + * + * Revision 1.15 2001/09/20 16:11:06 steudten + * + * Add casting for some string functions. + * + * Revision 1.14 2001/09/09 21:41:57 oes + * Fixing yet another silly bug + * * Revision 1.13 2001/09/06 14:05:59 oes * Fixed silly bug * @@ -109,7 +157,12 @@ const char pcrs_rcs[] = "$Id: pcrs.c,v 1.13 2001/09/06 14:05:59 oes Exp $"; *********************************************************************/ -#include +/* + * Include project.h just so that the right pcre.h gets + * included from there + */ +#include "project.h" + #include #include @@ -117,6 +170,14 @@ const char pcrs_rcs[] = "$Id: pcrs.c,v 1.13 2001/09/06 14:05:59 oes Exp $"; const char pcrs_h_rcs[] = PCRS_H_VERSION; +/* + * Internal prototypes + */ + +static int pcrs_parse_perl_options(const char *optstring, int *flags); +static pcrs_substitute *pcrs_compile_replacement(const char *replacement, int trivialflag, + int capturecount, int *errptr); + /********************************************************************* * * Function : pcrs_strerror @@ -148,6 +209,13 @@ const char *pcrs_strerror(const int error) case PCRE_ERROR_NOSUBSTRING: return "(pcre:) Fire in power supply"; case PCRE_ERROR_NOMATCH: return "(pcre:) Water in power supply"; +#ifdef PCRE_ERROR_MATCHLIMIT + /* + * Only reported by PCRE versions newer than our own. + */ + case PCRE_ERROR_MATCHLIMIT: return "(pcre:) Match limit reached"; +#endif /* def PCRE_ERROR_MATCHLIMIT */ + /* PCRS errors: */ case PCRS_ERR_NOMEM: return "(pcrs:) No memory"; case PCRS_ERR_CMDSYNTAX: return "(pcrs:) Syntax error while parsing command"; @@ -155,8 +223,13 @@ const char *pcrs_strerror(const int error) case PCRS_ERR_BADJOB: return "(pcrs:) Bad job - NULL job, pattern or substitute"; case PCRS_WARN_BADREF: return "(pcrs:) Backreference out of range"; - /* What's that? */ - default: return "Unknown error"; + /* + * XXX: With the exception of PCRE_ERROR_MATCHLIMIT we + * only catch PCRE errors that can happen with our internal + * version. If Privoxy is linked against a newer + * PCRE version all bets are off ... + */ + default: return "Unknown error. Privoxy out of sync with PCRE?"; } } /* error >= 0: No error */ @@ -184,7 +257,7 @@ const char *pcrs_strerror(const int error) * Returns : option integer suitable for pcre * *********************************************************************/ -int pcrs_parse_perl_options(const char *optstring, int *flags) +static int pcrs_parse_perl_options(const char *optstring, int *flags) { size_t i; int rc = 0; @@ -192,7 +265,7 @@ int pcrs_parse_perl_options(const char *optstring, int *flags) if (NULL == optstring) return 0; - for (i=0; i < strlen(optstring); i++) + for (i = 0; i < strlen(optstring); i++) { switch(optstring[i]) { @@ -204,7 +277,7 @@ int pcrs_parse_perl_options(const char *optstring, int *flags) case 's': rc |= PCRE_DOTALL; break; case 'x': rc |= PCRE_EXTENDED; break; case 'U': rc |= PCRE_UNGREEDY; break; - case 'T': *flags |= PCRS_TRIVIAL; break; + case 'T': *flags |= PCRS_TRIVIAL; break; default: break; } } @@ -237,9 +310,10 @@ int pcrs_parse_perl_options(const char *optstring, int *flags) * the reason. * *********************************************************************/ -pcrs_substitute *pcrs_compile_replacement(const char *replacement, int trivialflag, int capturecount, int *errptr) +static pcrs_substitute *pcrs_compile_replacement(const char *replacement, int trivialflag, int capturecount, int *errptr) { - int length, i, k, l, quoted; + int i, k, l, quoted; + size_t length; char *text; pcrs_substitute *r; @@ -280,7 +354,7 @@ pcrs_substitute *pcrs_compile_replacement(const char *replacement, int trivialfl if (trivialflag) { text = strncpy(text, replacement, length + 1); - k = length; + k = (int)length; } /* @@ -288,7 +362,7 @@ pcrs_substitute *pcrs_compile_replacement(const char *replacement, int trivialfl */ else { - while(i < length) + while (i < (int)length) { /* Quoting */ if (replacement[i] == '\\') @@ -302,7 +376,7 @@ pcrs_substitute *pcrs_compile_replacement(const char *replacement, int trivialfl { if (replacement[i+1] && strchr("tnrfae0", replacement[i+1])) { - switch(replacement[++i]) + switch (replacement[++i]) { case 't': text[k++] = '\t'; @@ -317,7 +391,7 @@ pcrs_substitute *pcrs_compile_replacement(const char *replacement, int trivialfl text[k++] = '\f'; break; case 'a': - text[k++] = '\a'; + text[k++] = 7; break; case 'e': text[k++] = 27; @@ -338,15 +412,15 @@ pcrs_substitute *pcrs_compile_replacement(const char *replacement, int trivialfl } /* Backreferences */ - if (replacement[i] == '$' && !quoted && i < length - 1) + if (replacement[i] == '$' && !quoted && i < (int)(length - 1)) { char *symbol, symbols[] = "'`+&"; - r->block_length[l] = k - r->block_offset[l]; + r->block_length[l] = (size_t)(k - r->block_offset[l]); /* Numerical backreferences */ - if (isdigit(replacement[i + 1])) + if (isdigit((int)replacement[i + 1])) { - while (i < length && isdigit(replacement[++i])) + while (i < (int)length && isdigit((int)replacement[++i])) { r->backref[l] = r->backref[l] * 10 + replacement[i] - 48; } @@ -406,7 +480,8 @@ plainchar: */ r->text = text; r->backrefs = l; - r->block_length[l] = k - r->block_offset[l]; + r->length = (size_t)k; + r->block_length[l] = (size_t)(k - r->block_offset[l]); return r; @@ -496,7 +571,8 @@ void pcrs_free_joblist(pcrs_job *joblist) *********************************************************************/ pcrs_job *pcrs_compile_command(const char *command, int *errptr) { - int i, k, l, limit, quoted = FALSE; + int i, k, l, quoted = FALSE; + size_t limit; char delimiter; char *tokens[4]; pcrs_job *newjob; @@ -519,7 +595,7 @@ pcrs_job *pcrs_compile_command(const char *command, int *errptr) tokens[l] = (char *) malloc(limit + 1); - for (i=0; i <= limit; i++) + for (i = 0; i <= (int)limit; i++) { if (command[i] == delimiter && !quoted) @@ -674,26 +750,30 @@ pcrs_job *pcrs_compile(const char *pattern, const char *substitute, const char * * the joblist to the subject. * The subject itself is left untouched, memory for the result * is malloc()ed and it is the caller's responsibility to free - * the result when it's no longer needed. + * the result when it's no longer needed. + * + * Note: For convenient string handling, a null byte is + * appended to the result. It does not count towards the + * result_length, though. + * * * Parameters : * 1 : joblist = the chained list of pcrs_jobs to be executed * 2 : subject = the subject string * 3 : subject_length = the subject's length - * INCLUDING the terminating zero, if string! * 4 : result = char** for returning the result * 5 : result_length = size_t* for returning the result's length * * Returns : On success, the number of substitutions that were made. * May be > 1 if job->flags contained PCRS_GLOBAL - * On failiure, the (negative) pcre error code describing the - * failiure, which may be translated to text using pcrs_strerror(). + * On failure, the (negative) pcre error code describing the + * failure, which may be translated to text using pcrs_strerror(). * *********************************************************************/ int pcrs_execute_list(pcrs_job *joblist, char *subject, size_t subject_length, char **result, size_t *result_length) { pcrs_job *job; - char *old, *new; + char *old, *new = NULL; int hits, total_hits; old = subject; @@ -733,28 +813,32 @@ int pcrs_execute_list(pcrs_job *joblist, char *subject, size_t subject_length, c * is malloc()ed and it is the caller's responsibility to free * the result when it's no longer needed. * + * Note: For convenient string handling, a null byte is + * appended to the result. It does not count towards the + * result_length, though. + * * Parameters : * 1 : job = the pcrs_job to be executed * 2 : subject = the subject (== original) string * 3 : subject_length = the subject's length - * INCLUDING the terminating zero, if string! * 4 : result = char** for returning the result * 5 : result_length = size_t* for returning the result's length * * Returns : On success, the number of substitutions that were made. * May be > 1 if job->flags contained PCRS_GLOBAL - * On failiure, the (negative) pcre error code describing the - * failiure, which may be translated to text using pcrs_strerror(). + * On failure, the (negative) pcre error code describing the + * failure, which may be translated to text using pcrs_strerror(). * *********************************************************************/ int pcrs_execute(pcrs_job *job, char *subject, size_t subject_length, char **result, size_t *result_length) { int offsets[3 * PCRS_MAX_SUBMATCHES], - offset, i, k, + offset, + i, k, matches_found, - newsize, submatches, max_matches = PCRS_MAX_MATCH_INIT; + size_t newsize; pcrs_match *matches, *dummy; char *result_offset; @@ -763,59 +847,59 @@ int pcrs_execute(pcrs_job *job, char *subject, size_t subject_length, char **res /* * Sanity check & memory allocation */ - if (job == NULL || job->pattern == NULL || job->substitute == NULL) + if (job == NULL || job->pattern == NULL || job->substitute == NULL || NULL == subject) { *result = NULL; return(PCRS_ERR_BADJOB); } - if (NULL == (matches = (pcrs_match *)malloc(max_matches * sizeof(pcrs_match)))) + if (NULL == (matches = (pcrs_match *)malloc((size_t)max_matches * sizeof(pcrs_match)))) { *result = NULL; return(PCRS_ERR_NOMEM); } - memset(matches, '\0', max_matches * sizeof(pcrs_match)); + memset(matches, '\0', (size_t)max_matches * sizeof(pcrs_match)); /* * Find the pattern and calculate the space * requirements for the result */ - newsize=subject_length; + newsize = subject_length; - while ((submatches = pcre_exec(job->pattern, job->hints, subject, subject_length, offset, 0, offsets, 3 * PCRS_MAX_SUBMATCHES)) > 0) + while ((submatches = pcre_exec(job->pattern, job->hints, subject, (int)subject_length, offset, 0, offsets, 3 * PCRS_MAX_SUBMATCHES)) > 0) { job->flags |= PCRS_SUCCESS; matches[i].submatches = submatches; - for (k=0; k < submatches; k++) + for (k = 0; k < submatches; k++) { matches[i].submatch_offset[k] = offsets[2 * k]; /* Note: Non-found optional submatches have length -1-(-1)==0 */ - matches[i].submatch_length[k] = offsets[2 * k + 1] - offsets[2 * k]; + matches[i].submatch_length[k] = (size_t)(offsets[2 * k + 1] - offsets[2 * k]); /* reserve mem for each submatch as often as it is ref'd */ - newsize += matches[i].submatch_length[k] * job->substitute->backref_count[k]; + newsize += matches[i].submatch_length[k] * (size_t)job->substitute->backref_count[k]; } /* plus replacement text size minus match text size */ - newsize += strlen(job->substitute->text) - matches[i].submatch_length[0]; + newsize += job->substitute->length - matches[i].submatch_length[0]; /* chunk before match */ matches[i].submatch_offset[PCRS_MAX_SUBMATCHES] = 0; - matches[i].submatch_length[PCRS_MAX_SUBMATCHES] = offsets[0]; - newsize += offsets[0] * job->substitute->backref_count[PCRS_MAX_SUBMATCHES]; + matches[i].submatch_length[PCRS_MAX_SUBMATCHES] = (size_t)offsets[0]; + newsize += (size_t)offsets[0] * (size_t)job->substitute->backref_count[PCRS_MAX_SUBMATCHES]; /* chunk after match */ matches[i].submatch_offset[PCRS_MAX_SUBMATCHES + 1] = offsets[1]; - matches[i].submatch_length[PCRS_MAX_SUBMATCHES + 1] = subject_length - offsets[1] - 1; - newsize += (subject_length - offsets[1]) * job->substitute->backref_count[PCRS_MAX_SUBMATCHES + 1]; + matches[i].submatch_length[PCRS_MAX_SUBMATCHES + 1] = subject_length - (size_t)offsets[1] - 1; + newsize += (subject_length - (size_t)offsets[1]) * (size_t)job->substitute->backref_count[PCRS_MAX_SUBMATCHES + 1]; /* Storage for matches exhausted? -> Extend! */ if (++i >= max_matches) { - max_matches *= PCRS_MAX_MATCH_GROW; - if (NULL == (dummy = (pcrs_match *)realloc(matches, max_matches * sizeof(pcrs_match)))) + max_matches = (int)(max_matches * PCRS_MAX_MATCH_GROW); + if (NULL == (dummy = (pcrs_match *)realloc(matches, (size_t)max_matches * sizeof(pcrs_match)))) { free(matches); *result = NULL; @@ -829,7 +913,7 @@ int pcrs_execute(pcrs_job *job, char *subject, size_t subject_length, char **res /* Don't loop on empty matches */ if (offsets[1] == offset) - if (offset < subject_length) + if ((size_t)offset < subject_length) offset++; else break; @@ -847,13 +931,18 @@ int pcrs_execute(pcrs_job *job, char *subject, size_t subject_length, char **res /* - * Get memory for the result + * Get memory for the result (must be freed by caller!) + * and append terminating null byte. */ - if ((*result = (char *)malloc(newsize)) == NULL) /* must be free()d by caller */ + if ((*result = (char *)malloc(newsize + 1)) == NULL) { free(matches); return PCRS_ERR_NOMEM; } + else + { + (*result)[newsize] = '\0'; + } /* @@ -862,14 +951,14 @@ int pcrs_execute(pcrs_job *job, char *subject, size_t subject_length, char **res offset = 0; result_offset = *result; - for (i=0; i < matches_found; i++) + for (i = 0; i < matches_found; i++) { /* copy the chunk preceding the match */ - memcpy(result_offset, subject + offset, matches[i].submatch_offset[0] - offset); + memcpy(result_offset, subject + offset, (size_t)(matches[i].submatch_offset[0] - offset)); result_offset += matches[i].submatch_offset[0] - offset; /* For every segment of the substitute.. */ - for (k=0; k <= job->substitute->backrefs; k++) + for (k = 0; k <= job->substitute->backrefs; k++) { /* ...copy its text.. */ memcpy(result_offset, job->substitute->text + job->substitute->block_offset[k], job->substitute->block_length[k]); @@ -879,7 +968,9 @@ int pcrs_execute(pcrs_job *job, char *subject, size_t subject_length, char **res if (k != job->substitute->backrefs /* ..in legal range.. */ && job->substitute->backref[k] < PCRS_MAX_SUBMATCHES + 2 - /* ..and referencing a nonempty match.. */ + /* ..and referencing a real submatch.. */ + && job->substitute->backref[k] < matches[i].submatches + /* ..that is nonempty.. */ && matches[i].submatch_length[job->substitute->backref[k]] > 0) { /* ..copy the submatch that is ref'd. */ @@ -891,11 +982,11 @@ int pcrs_execute(pcrs_job *job, char *subject, size_t subject_length, char **res result_offset += matches[i].submatch_length[job->substitute->backref[k]]; } } - offset = matches[i].submatch_offset[0] + matches[i].submatch_length[0]; + offset = matches[i].submatch_offset[0] + (int)matches[i].submatch_length[0]; } /* Copy the rest. */ - memcpy(result_offset, subject + offset, subject_length - offset); + memcpy(result_offset, subject + offset, subject_length - (size_t)offset); *result_length = newsize; free(matches);