--- code/trunk/pcretest.c 2007/04/02 13:32:07 144 +++ code/trunk/pcretest.c 2007/09/11 12:57:06 236 @@ -36,6 +36,10 @@ */ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + #include #include #include @@ -67,13 +71,16 @@ #endif -#define PCRE_SPY /* For Win32 build, import data, not export */ - -/* We include pcre_internal.h because we need the internal info for displaying -the results of pcre_study() and we also need to know about the internal -macros, structures, and other internal data values; pcretest has "inside -information" compared to a program that strictly follows the PCRE API. */ +/* We have to include pcre_internal.h because we need the internal info for +displaying the results of pcre_study() and we also need to know about the +internal macros, structures, and other internal data values; pcretest has +"inside information" compared to a program that strictly follows the PCRE API. + +Although pcre_internal.h does itself include pcre.h, we explicitly include it +here before pcre_internal.h so that the PCRE_EXP_xxx macros get set +appropriately for an application, not for building PCRE. */ +#include "pcre.h" #include "pcre_internal.h" /* We need access to the data tables that PCRE uses. So as not to have to keep @@ -149,6 +156,7 @@ static int callout_extra; static int callout_fail_count; static int callout_fail_id; +static int debug_lengths; static int first_callout; static int locale_set = 0; static int show_malloc; @@ -656,11 +664,38 @@ /************************************************* +* Case-independent strncmp() function * +*************************************************/ + +/* +Arguments: + s first string + t second string + n number of characters to compare + +Returns: < 0, = 0, or > 0, according to the comparison +*/ + +static int +strncmpic(uschar *s, uschar *t, int n) +{ +while (n--) + { + int c = tolower(*s++) - tolower(*t++); + if (c) return c; + } +return 0; +} + + + +/************************************************* * Check newline indicator * *************************************************/ /* This is used both at compile and run-time to check for escapes, where -xxx is LF, CR, CRLF, or ANY. Print a message and return 0 if there is no match. +xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is +no match. Arguments: p points after the leading '<' @@ -672,10 +707,13 @@ static int check_newline(uschar *p, FILE *f) { -if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR; -if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF; -if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF; -if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY; +if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR; +if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF; +if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF; +if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF; +if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY; +if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF; +if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE; fprintf(f, "Unknown newline type at: <%s\n", p); return 0; } @@ -847,7 +885,11 @@ (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); printf(" Newline sequence is %s\n", (rc == '\r')? "CR" : (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" : + (rc == -2)? "ANYCRLF" : (rc == -1)? "ANY" : "???"); + (void)pcre_config(PCRE_CONFIG_BSR, &rc); + printf(" \\R matches %s\n", rc? "CR, LF, or CRLF only" : + "all Unicode newlines"); (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); printf(" Internal link size = %d\n", rc); (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc); @@ -884,7 +926,7 @@ if (offsets == NULL) { printf("** Failed to get %d bytes of memory for offsets vector\n", - size_offsets_max * sizeof(int)); + (int)(size_offsets_max * sizeof(int))); yield = 1; goto EXIT; } @@ -944,7 +986,6 @@ size_t size, regex_gotten_store; int do_study = 0; int do_debug = debug; - int debug_lengths = 1; int do_G = 0; int do_g = 0; int do_showinfo = showinfo; @@ -953,6 +994,7 @@ int erroroffset, len, delimiter, poffset; use_utf8 = 0; + debug_lengths = 1; if (infile == stdin) printf(" re> "); if (extend_inputline(infile, buffer) == NULL) break; @@ -1312,6 +1354,7 @@ rre->magic_number = byteflip(rre->magic_number, sizeof(rre->magic_number)); rre->size = byteflip(rre->size, sizeof(rre->size)); rre->options = byteflip(rre->options, sizeof(rre->options)); + rre->flags = byteflip(rre->flags, sizeof(rre->flags)); rre->top_bracket = byteflip(rre->top_bracket, sizeof(rre->top_bracket)); rre->top_backref = byteflip(rre->top_backref, sizeof(rre->top_backref)); rre->first_byte = byteflip(rre->first_byte, sizeof(rre->first_byte)); @@ -1346,7 +1389,8 @@ #if !defined NOINFOCHECK int old_first_char, old_options, old_count; #endif - int count, backrefmax, first_char, need_char; + int count, backrefmax, first_char, need_char, okpartial, jchanged, + hascrorlf; int nameentrysize, namecount; const uschar *nametable; @@ -1359,6 +1403,9 @@ new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize); new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount); new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable); + new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial); + new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged); + new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf); #if !defined NOINFOCHECK old_count = pcre_info(re, &old_options, &old_first_char); @@ -1400,26 +1447,22 @@ } } - /* The NOPARTIAL bit is a private bit in the options, so we have - to fish it out via out back door */ + if (!okpartial) fprintf(outfile, "Partial matching not supported\n"); + if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n"); all_options = ((real_pcre *)re)->options; - if (do_flip) - { - all_options = byteflip(all_options, sizeof(all_options)); - } - - if ((all_options & PCRE_NOPARTIAL) != 0) - fprintf(outfile, "Partial matching not supported\n"); + if (do_flip) all_options = byteflip(all_options, sizeof(all_options)); if (get_options == 0) fprintf(outfile, "No options\n"); - else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s\n", + else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n", ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "", ((get_options & PCRE_CASELESS) != 0)? " caseless" : "", ((get_options & PCRE_EXTENDED) != 0)? " extended" : "", ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "", ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "", ((get_options & PCRE_DOTALL) != 0)? " dotall" : "", + ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "", + ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "", ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "", ((get_options & PCRE_EXTRA) != 0)? " extra" : "", ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "", @@ -1428,6 +1471,8 @@ ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "", ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : ""); + if (jchanged) fprintf(outfile, "Duplicate name status changes\n"); + switch (get_options & PCRE_NEWLINE_BITS) { case PCRE_NEWLINE_CR: @@ -1442,6 +1487,10 @@ fprintf(outfile, "Forced newline sequence: CRLF\n"); break; + case PCRE_NEWLINE_ANYCRLF: + fprintf(outfile, "Forced newline sequence: ANYCRLF\n"); + break; + case PCRE_NEWLINE_ANY: fprintf(outfile, "Forced newline sequence: ANY\n"); break; @@ -1591,7 +1640,7 @@ for (;;) { uschar *q; - uschar *bptr = dbuffer; + uschar *bptr; int *use_offsets = offsets; int use_size_offsets = size_offsets; int callout_data = 0; @@ -1647,7 +1696,7 @@ p = buffer; while (isspace(*p)) p++; - q = dbuffer; + bptr = q = dbuffer; while ((c = *p++) != 0) { int i = 0; @@ -1842,7 +1891,7 @@ if (offsets == NULL) { printf("** Failed to get %d bytes of memory for offsets vector\n", - size_offsets_max * sizeof(int)); + (int)(size_offsets_max * sizeof(int))); yield = 1; goto EXIT; } @@ -2215,11 +2264,11 @@ to advance the start offset, and continue. We won't be at the end of the string - that was checked before setting g_notempty. - Complication arises in the case when the newline option is "any". - If the previous match was at the end of a line terminated by CRLF, an - advance of one character just passes the \r, whereas we should prefer the - longer newline sequence, as does the code in pcre_exec(). Fudge the - offset value to achieve this. + Complication arises in the case when the newline option is "any" or + "anycrlf". If the previous match was at the end of a line terminated by + CRLF, an advance of one character just passes the \r, whereas we should + prefer the longer newline sequence, as does the code in pcre_exec(). + Fudge the offset value to achieve this. Otherwise, in the case of UTF-8 matching, the advance must be one character, not one byte. */ @@ -2229,9 +2278,21 @@ if (g_notempty != 0) { int onechar = 1; + unsigned int obits = ((real_pcre *)re)->options; use_offsets[0] = start_offset; - if ((((real_pcre *)re)->options & PCRE_NEWLINE_BITS) == - PCRE_NEWLINE_ANY && + if ((obits & PCRE_NEWLINE_BITS) == 0) + { + int d; + (void)pcre_config(PCRE_CONFIG_NEWLINE, &d); + obits = (d == '\r')? PCRE_NEWLINE_CR : + (d == '\n')? PCRE_NEWLINE_LF : + (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF : + (d == -2)? PCRE_NEWLINE_ANYCRLF : + (d == -1)? PCRE_NEWLINE_ANY : 0; + } + if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY || + (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF) + && start_offset < len - 1 && bptr[start_offset] == '\r' && bptr[start_offset+1] == '\n')