-/* $Cambridge: exim/src/src/pcre/pcretest.c,v 1.3 2005/08/08 10:22:14 ph10 Exp $ */
+/* $Cambridge: exim/src/src/pcre/pcretest.c,v 1.6 2007/01/23 15:08:45 ph10 Exp $ */
/*************************************************
* PCRE testing program *
#include <locale.h>
#include <errno.h>
+
+/* A number of things vary for Windows builds. Originally, pcretest opened its
+input and output without "b"; then I was told that "b" was needed in some
+environments, so it was added for release 5.0 to both the input and output. (It
+makes no difference on Unix-like systems.) Later I was told that it is wrong
+for the input on Windows. I've now abstracted the modes into two macros that
+are set here, to make it easier to fiddle with them, and removed "b" from the
+input mode under Windows. */
+
+#if defined(_WIN32) || defined(WIN32)
+#include <io.h> /* For _setmode() */
+#include <fcntl.h> /* For _O_BINARY */
+#define INPUT_MODE "r"
+#define OUTPUT_MODE "wb"
+
+#else
+#include <sys/time.h> /* These two includes are needed */
+#include <sys/resource.h> /* for setrlimit(). */
+#define INPUT_MODE "rb"
+#define OUTPUT_MODE "wb"
+#endif
+
+
#define PCRE_SPY /* For Win32 build, import data, not export */
-/* We need the internal info for displaying the results of pcre_study() and
-other internal data; pcretest also uses some of the fixed tables, and generally
-has "inside information" compared to a program that strictly follows the PCRE
-API. */
+/* 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. */
#include "pcre_internal.h"
+/* We need access to the data tables that PCRE uses. So as not to have to keep
+two copies, we include the source file here, changing the names of the external
+symbols to prevent clashes. */
+
+#define _pcre_utf8_table1 utf8_table1
+#define _pcre_utf8_table1_size utf8_table1_size
+#define _pcre_utf8_table2 utf8_table2
+#define _pcre_utf8_table3 utf8_table3
+#define _pcre_utf8_table4 utf8_table4
+#define _pcre_utt utt
+#define _pcre_utt_size utt_size
+#define _pcre_OP_lengths OP_lengths
+
+#include "pcre_tables.c"
+
+/* We also need the pcre_printint() function for printing out compiled
+patterns. This function is in a separate file so that it can be included in
+pcre_compile.c when that module is compiled with debugging enabled.
+
+The definition of the macro PRINTABLE, which determines whether to print an
+output character as-is or as a hex value when showing compiled patterns, is
+contained in this file. We uses it here also, in cases when the locale has not
+been explicitly changed, so as to get consistent output from systems that
+differ in their output from isprint() even in the "C" locale. */
+
+#include "pcre_printint.src"
+
+#define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c))
+
/* It is possible to compile this test program without including support for
testing the POSIX interface, though this is not available via the standard
function (define NOINFOCHECK). */
+/* Other parameters */
+
#ifndef CLOCKS_PER_SEC
#ifdef CLK_TCK
#define CLOCKS_PER_SEC CLK_TCK
#endif
#endif
-#define LOOPREPEAT 500000
+/* This is the default loop count for timing. */
-#define BUFFER_SIZE 30000
-#define PBUFFER_SIZE BUFFER_SIZE
-#define DBUFFER_SIZE BUFFER_SIZE
+#define LOOPREPEAT 500000
+/* Static variables */
static FILE *outfile;
static int log_store = 0;
static int callout_fail_count;
static int callout_fail_id;
static int first_callout;
+static int locale_set = 0;
static int show_malloc;
static int use_utf8;
static size_t gotten_store;
+/* The buffers grow automatically if very long input lines are encountered. */
+
+static int buffer_size = 50000;
+static uschar *buffer = NULL;
+static uschar *dbuffer = NULL;
static uschar *pbuffer = NULL;
+/*************************************************
+* Read or extend an input line *
+*************************************************/
+
+/* Input lines are read into buffer, but both patterns and data lines can be
+continued over multiple input lines. In addition, if the buffer fills up, we
+want to automatically expand it so as to be able to handle extremely large
+lines that are needed for certain stress tests. When the input buffer is
+expanded, the other two buffers must also be expanded likewise, and the
+contents of pbuffer, which are a copy of the input for callouts, must be
+preserved (for when expansion happens for a data line). This is not the most
+optimal way of handling this, but hey, this is just a test program!
+
+Arguments:
+ f the file to read
+ start where in buffer to start (this *must* be within buffer)
+
+Returns: pointer to the start of new data
+ could be a copy of start, or could be moved
+ NULL if no data read and EOF reached
+*/
+
+static uschar *
+extend_inputline(FILE *f, uschar *start)
+{
+uschar *here = start;
+
+for (;;)
+ {
+ int rlen = buffer_size - (here - buffer);
+
+ if (rlen > 1000)
+ {
+ int dlen;
+ if (fgets((char *)here, rlen, f) == NULL)
+ return (here == start)? NULL : start;
+ dlen = (int)strlen((char *)here);
+ if (dlen > 0 && here[dlen - 1] == '\n') return start;
+ here += dlen;
+ }
+
+ else
+ {
+ int new_buffer_size = 2*buffer_size;
+ uschar *new_buffer = (unsigned char *)malloc(new_buffer_size);
+ uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size);
+ uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size);
+
+ if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL)
+ {
+ fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size);
+ exit(1);
+ }
+
+ memcpy(new_buffer, buffer, buffer_size);
+ memcpy(new_pbuffer, pbuffer, buffer_size);
+
+ buffer_size = new_buffer_size;
+
+ start = new_buffer + (start - buffer);
+ here = new_buffer + (here - buffer);
+
+ free(buffer);
+ free(dbuffer);
+ free(pbuffer);
+
+ buffer = new_buffer;
+ dbuffer = new_dbuffer;
+ pbuffer = new_pbuffer;
+ }
+ }
+
+return NULL; /* Control never gets here */
+}
+
+
+
+
+
+
+
/*************************************************
* Read number from string *
*************************************************/
/* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess
around with conditional compilation, just do the job by hand. It is only used
-for unpicking the -o argument, so just keep it simple.
+for unpicking arguments, so just keep it simple.
Arguments:
str string to be converted
and returns the value of the character.
Argument:
- buffer a pointer to the byte vector
- vptr a pointer to an int to receive the value
+ utf8bytes a pointer to the byte vector
+ vptr a pointer to an int to receive the value
-Returns: > 0 => the number of bytes consumed
- -6 to 0 => malformed UTF-8 character at offset = (-return)
+Returns: > 0 => the number of bytes consumed
+ -6 to 0 => malformed UTF-8 character at offset = (-return)
*/
#if !defined NOUTF8
static int
-utf82ord(unsigned char *buffer, int *vptr)
+utf82ord(unsigned char *utf8bytes, int *vptr)
{
-int c = *buffer++;
+int c = *utf8bytes++;
int d = c;
int i, j, s;
/* i now has a value in the range 1-5 */
s = 6*i;
-d = (c & _pcre_utf8_table3[i]) << s;
+d = (c & utf8_table3[i]) << s;
for (j = 0; j < i; j++)
{
- c = *buffer++;
+ c = *utf8bytes++;
if ((c & 0xc0) != 0x80) return -(j+1);
s -= 6;
d |= (c & 0x3f) << s;
/* Check that encoding was the correct unique one */
-for (j = 0; j < _pcre_utf8_table1_size; j++)
- if (d <= _pcre_utf8_table1[j]) break;
+for (j = 0; j < utf8_table1_size; j++)
+ if (d <= utf8_table1[j]) break;
if (j != i) return -(i+1);
/* Valid value */
+/*************************************************
+* Convert character value to UTF-8 *
+*************************************************/
+
+/* This function takes an integer value in the range 0 - 0x7fffffff
+and encodes it as a UTF-8 character in 0 to 6 bytes.
+
+Arguments:
+ cvalue the character value
+ utf8bytes pointer to buffer for result - at least 6 bytes long
+
+Returns: number of characters placed in the buffer
+*/
+
+#if !defined NOUTF8
+
+static int
+ord2utf8(int cvalue, uschar *utf8bytes)
+{
+register int i, j;
+for (i = 0; i < utf8_table1_size; i++)
+ if (cvalue <= utf8_table1[i]) break;
+utf8bytes += i;
+for (j = i; j > 0; j--)
+ {
+ *utf8bytes-- = 0x80 | (cvalue & 0x3f);
+ cvalue >>= 6;
+ }
+*utf8bytes = utf8_table2[i] | cvalue;
+return i + 1;
+}
+
+#endif
+
+
+
/*************************************************
* Print character string *
*************************************************/
static int pchars(unsigned char *p, int length, FILE *f)
{
-int c;
+int c = 0;
int yield = 0;
while (length-- > 0)
{
length -= rc - 1;
p += rc;
- if (c < 256 && isprint(c))
+ if (PRINTHEX(c))
{
if (f != NULL) fprintf(f, "%c", c);
yield++;
}
else
{
- int n;
- if (f != NULL) fprintf(f, "\\x{%02x}%n", c, &n);
- yield += n;
+ int n = 4;
+ if (f != NULL) fprintf(f, "\\x{%02x}", c);
+ yield += (n <= 0x000000ff)? 2 :
+ (n <= 0x00000fff)? 3 :
+ (n <= 0x0000ffff)? 4 :
+ (n <= 0x000fffff)? 5 : 6;
}
continue;
}
/* Not UTF-8, or malformed UTF-8 */
- if (isprint(c = *(p++)))
+ c = *p++;
+ if (PRINTHEX(c))
{
if (f != NULL) fprintf(f, "%c", c);
yield++;
* Byte flipping function *
*************************************************/
-static long int
-byteflip(long int value, int n)
+static unsigned long int
+byteflip(unsigned long int value, int n)
{
if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8);
return ((value & 0x000000ff) << 24) |
+/*************************************************
+* Check match or recursion limit *
+*************************************************/
+
+static int
+check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len,
+ int start_offset, int options, int *use_offsets, int use_size_offsets,
+ int flag, unsigned long int *limit, int errnumber, const char *msg)
+{
+int count;
+int min = 0;
+int mid = 64;
+int max = -1;
+
+extra->flags |= flag;
+
+for (;;)
+ {
+ *limit = mid;
+
+ count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options,
+ use_offsets, use_size_offsets);
+
+ if (count == errnumber)
+ {
+ /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
+ min = mid;
+ mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
+ }
+
+ else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
+ count == PCRE_ERROR_PARTIAL)
+ {
+ if (mid == min + 1)
+ {
+ fprintf(outfile, "Minimum %s limit = %d\n", msg, mid);
+ break;
+ }
+ /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */
+ max = mid;
+ mid = (min + mid)/2;
+ }
+ else break; /* Some other error */
+ }
+
+extra->flags &= ~flag;
+return count;
+}
+
+
+
+/*************************************************
+* Check newline indicator *
+*************************************************/
+
+/* This is used both at compile and run-time to check for <xxx> escapes, where
+xxx is LF, CR, CRLF, or ANY. Print a message and return 0 if there is no match.
+
+Arguments:
+ p points after the leading '<'
+ f file for error message
+
+Returns: appropriate PCRE_NEWLINE_xxx flags, or 0
+*/
+
+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;
+fprintf(f, "Unknown newline type at: <%s\n", p);
+return 0;
+}
+
+
+
+/*************************************************
+* Usage function *
+*************************************************/
+
+static void
+usage(void)
+{
+printf("Usage: pcretest [options] [<input> [<output>]]\n");
+printf(" -b show compiled code (bytecode)\n");
+printf(" -C show PCRE compile-time options and exit\n");
+printf(" -d debug: show compiled code and information (-b and -i)\n");
+#if !defined NODFA
+printf(" -dfa force DFA matching for all subjects\n");
+#endif
+printf(" -help show usage information\n");
+printf(" -i show information about compiled patterns\n"
+ " -m output memory used information\n"
+ " -o <n> set size of offsets vector to <n>\n");
+#if !defined NOPOSIX
+printf(" -p use POSIX interface\n");
+#endif
+printf(" -q quiet: do not output PCRE version number at start\n");
+printf(" -S <n> set stack size to <n> megabytes\n");
+printf(" -s output store (memory) used information\n"
+ " -t time compilation and execution\n");
+printf(" -t <n> time compilation and execution, repeating <n> times\n");
+printf(" -tm time execution (matching) only\n");
+printf(" -tm <n> time execution (matching) only, repeating <n> times\n");
+}
+
+
+
/*************************************************
* Main Program *
*************************************************/
int study_options = 0;
int op = 1;
int timeit = 0;
+int timeitm = 0;
int showinfo = 0;
int showstore = 0;
+int quiet = 0;
int size_offsets = 45;
int size_offsets_max;
int *offsets = NULL;
int done = 0;
int all_use_dfa = 0;
int yield = 0;
+int stack_size;
-unsigned char *buffer;
-unsigned char *dbuffer;
+/* These vectors store, end-to-end, a list of captured substring names. Assume
+that 1024 is plenty long enough for the few names we'll be testing. */
-/* Get buffers from malloc() so that Electric Fence will check their misuse
-when I am debugging. */
+uschar copynames[1024];
+uschar getnames[1024];
-buffer = (unsigned char *)malloc(BUFFER_SIZE);
-dbuffer = (unsigned char *)malloc(DBUFFER_SIZE);
-pbuffer = (unsigned char *)malloc(PBUFFER_SIZE);
+uschar *copynamesptr;
+uschar *getnamesptr;
-/* The outfile variable is static so that new_malloc can use it. The _setmode()
-stuff is some magic that I don't understand, but which apparently does good
-things in Windows. It's related to line terminations. */
+/* Get buffers from malloc() so that Electric Fence will check their misuse
+when I am debugging. They grow automatically when very long lines are read. */
-#if defined(_WIN32) || defined(WIN32)
-_setmode( _fileno( stdout ), 0x8000 );
-#endif /* defined(_WIN32) || defined(WIN32) */
+buffer = (unsigned char *)malloc(buffer_size);
+dbuffer = (unsigned char *)malloc(buffer_size);
+pbuffer = (unsigned char *)malloc(buffer_size);
+
+/* The outfile variable is static so that new_malloc can use it. */
outfile = stdout;
+/* The following _setmode() stuff is some Windows magic that tells its runtime
+library to translate CRLF into a single LF character. At least, that's what
+I've been told: never having used Windows I take this all on trust. Originally
+it set 0x8000, but then I was advised that _O_BINARY was better. */
+
+#if defined(_WIN32) || defined(WIN32)
+_setmode( _fileno( stdout ), _O_BINARY );
+#endif
+
/* Scan options */
while (argc > 1 && argv[op][0] == '-')
if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0)
showstore = 1;
- else if (strcmp(argv[op], "-t") == 0) timeit = 1;
+ else if (strcmp(argv[op], "-q") == 0) quiet = 1;
+ else if (strcmp(argv[op], "-b") == 0) debug = 1;
else if (strcmp(argv[op], "-i") == 0) showinfo = 1;
else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1;
#if !defined NODFA
op++;
argc--;
}
+ else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0)
+ {
+ int both = argv[op][2] == 0;
+ int temp;
+ if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr),
+ *endptr == 0))
+ {
+ timeitm = temp;
+ op++;
+ argc--;
+ }
+ else timeitm = LOOPREPEAT;
+ if (both) timeit = timeitm;
+ }
+ else if (strcmp(argv[op], "-S") == 0 && argc > 2 &&
+ ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)),
+ *endptr == 0))
+ {
+#if defined(_WIN32) || defined(WIN32)
+ printf("PCRE: -S not supported on this OS\n");
+ exit(1);
+#else
+ int rc;
+ struct rlimit rlim;
+ getrlimit(RLIMIT_STACK, &rlim);
+ rlim.rlim_cur = stack_size * 1024 * 1024;
+ rc = setrlimit(RLIMIT_STACK, &rlim);
+ if (rc != 0)
+ {
+ printf("PCRE: setrlimit() failed with error %d\n", rc);
+ exit(1);
+ }
+ op++;
+ argc--;
+#endif
+ }
#if !defined NOPOSIX
else if (strcmp(argv[op], "-p") == 0) posix = 1;
#endif
(void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc);
printf(" %sUnicode properties support\n", rc? "" : "No ");
(void)pcre_config(PCRE_CONFIG_NEWLINE, &rc);
- printf(" Newline character is %s\n", (rc == '\r')? "CR" : "LF");
+ printf(" Newline sequence is %s\n", (rc == '\r')? "CR" :
+ (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" :
+ (rc == -1)? "ANY" : "???");
(void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc);
printf(" Internal link size = %d\n", rc);
(void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc);
printf(" POSIX malloc threshold = %d\n", rc);
(void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc);
printf(" Default match limit = %d\n", rc);
+ (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc);
+ printf(" Default recursion depth limit = %d\n", rc);
(void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc);
printf(" Match recursion uses %s\n", rc? "stack" : "heap");
exit(0);
}
+ else if (strcmp(argv[op], "-help") == 0 ||
+ strcmp(argv[op], "--help") == 0)
+ {
+ usage();
+ goto EXIT;
+ }
else
{
printf("** Unknown or malformed option %s\n", argv[op]);
- printf("Usage: pcretest [-d] [-i] [-o <n>] [-p] [-s] [-t] [<input> [<output>]]\n");
- printf(" -C show PCRE compile-time options and exit\n");
- printf(" -d debug: show compiled code; implies -i\n");
-#if !defined NODFA
- printf(" -dfa force DFA matching for all subjects\n");
-#endif
- printf(" -i show information about compiled pattern\n"
- " -m output memory used information\n"
- " -o <n> set size of offsets vector to <n>\n");
-#if !defined NOPOSIX
- printf(" -p use POSIX interface\n");
-#endif
- printf(" -s output store (memory) used information\n"
- " -t time compilation and execution\n");
+ usage();
yield = 1;
goto EXIT;
}
if (argc > 1)
{
- infile = fopen(argv[op], "rb");
+ infile = fopen(argv[op], INPUT_MODE);
if (infile == NULL)
{
printf("** Failed to open %s\n", argv[op]);
if (argc > 2)
{
- outfile = fopen(argv[op+1], "wb");
+ outfile = fopen(argv[op+1], OUTPUT_MODE);
if (outfile == NULL)
{
printf("** Failed to open %s\n", argv[op+1]);
pcre_stack_malloc = stack_malloc;
pcre_stack_free = stack_free;
-/* Heading line, then prompt for first regex if stdin */
+/* Heading line unless quiet, then prompt for first regex if stdin */
-fprintf(outfile, "PCRE version %s\n\n", pcre_version());
+if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version());
/* Main loop */
int do_showinfo = showinfo;
int do_showrest = 0;
int do_flip = 0;
- int erroroffset, len, delimiter;
+ int erroroffset, len, delimiter, poffset;
use_utf8 = 0;
if (infile == stdin) printf(" re> ");
- if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL) break;
+ if (extend_inputline(infile, buffer) == NULL) break;
if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
fflush(outfile);
if (*p == '<' && strchr((char *)(p+1), '<') == NULL)
{
- unsigned long int magic;
+ unsigned long int magic, get_options;
uschar sbuf[8];
FILE *f;
/* Need to know if UTF-8 for printing data strings */
- new_info(re, NULL, PCRE_INFO_OPTIONS, &options);
- use_utf8 = (options & PCRE_UTF8) != 0;
+ new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
+ use_utf8 = (get_options & PCRE_UTF8) != 0;
/* Now see if there is any following study data */
}
pp = p;
+ poffset = p - buffer;
for(;;)
{
pp++;
}
if (*pp != 0) break;
-
- len = BUFFER_SIZE - (pp - buffer);
- if (len < 256)
- {
- fprintf(outfile, "** Expression too long - missing delimiter?\n");
- goto SKIP_DATA;
- }
-
if (infile == stdin) printf(" > ");
- if (fgets((char *)pp, len, infile) == NULL)
+ if ((pp = extend_inputline(infile, pp)) == NULL)
{
fprintf(outfile, "** Unexpected EOF\n");
done = 1;
if (infile != stdin) fprintf(outfile, "%s", (char *)pp);
}
+ /* The buffer may have moved while being extended; reset the start of data
+ pointer to the correct relative point in the buffer. */
+
+ p = buffer + poffset;
+
/* If the first character after the delimiter is backslash, make
the pattern end with backslash. This is purely to provide a way
of testing for the error message when a pattern ends with backslash. */
case '+': do_showrest = 1; break;
case 'A': options |= PCRE_ANCHORED; break;
+ case 'B': do_debug = 1; break;
case 'C': options |= PCRE_AUTO_CALLOUT; break;
case 'D': do_debug = do_showinfo = 1; break;
case 'E': options |= PCRE_DOLLAR_ENDONLY; break;
case 'F': do_flip = 1; break;
case 'G': do_G = 1; break;
case 'I': do_showinfo = 1; break;
+ case 'J': options |= PCRE_DUPNAMES; break;
case 'M': log_store = 1; break;
case 'N': options |= PCRE_NO_AUTO_CAPTURE; break;
case 'L':
ppp = pp;
- /* The '\r' test here is so that it works on Windows */
- while (*ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
+ /* The '\r' test here is so that it works on Windows. */
+ /* The '0' test is just in case this is an unterminated line. */
+ while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++;
*ppp = 0;
if (setlocale(LC_CTYPE, (const char *)pp) == NULL)
{
fprintf(outfile, "** Failed to set locale \"%s\"\n", pp);
goto SKIP_DATA;
}
+ locale_set = 1;
tables = pcre_maketables();
pp = ppp;
break;
*pp = 0;
break;
+ case '<':
+ {
+ int x = check_newline(pp, outfile);
+ if (x == 0) goto SKIP_DATA;
+ options |= x;
+ while (*pp++ != '>');
+ }
+ break;
+
case '\r': /* So that it works in Windows */
case '\n':
case ' ':
if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE;
if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE;
if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL;
+ if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB;
+ if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8;
+
rc = regcomp(&preg, (char *)p, cflags);
/* Compilation failed; go back for another re, skipping to blank line
if (rc != 0)
{
- (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);
+ (void)regerror(rc, &preg, (char *)buffer, buffer_size);
fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer);
goto SKIP_DATA;
}
#endif /* !defined NOPOSIX */
{
- if (timeit)
+ if (timeit > 0)
{
register int i;
clock_t time_taken;
clock_t start_time = clock();
- for (i = 0; i < LOOPREPEAT; i++)
+ for (i = 0; i < timeit; i++)
{
re = pcre_compile((char *)p, options, &error, &erroroffset, tables);
if (re != NULL) free(re);
}
time_taken = clock() - start_time;
- fprintf(outfile, "Compile time %.3f milliseconds\n",
- (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /
+ fprintf(outfile, "Compile time %.4f milliseconds\n",
+ (((double)time_taken * 1000.0) / (double)timeit) /
(double)CLOCKS_PER_SEC);
}
{
for (;;)
{
- if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)
+ if (extend_inputline(infile, buffer) == NULL)
{
done = 1;
goto CONTINUE;
if (do_study)
{
- if (timeit)
+ if (timeit > 0)
{
register int i;
clock_t time_taken;
clock_t start_time = clock();
- for (i = 0; i < LOOPREPEAT; i++)
+ for (i = 0; i < timeit; i++)
extra = pcre_study(re, study_options, &error);
time_taken = clock() - start_time;
if (extra != NULL) free(extra);
- fprintf(outfile, " Study time %.3f milliseconds\n",
- (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /
+ fprintf(outfile, " Study time %.4f milliseconds\n",
+ (((double)time_taken * 1000.0) / (double)timeit) /
(double)CLOCKS_PER_SEC);
}
extra = pcre_study(re, study_options, &error);
SHOW_INFO:
+ if (do_debug)
+ {
+ fprintf(outfile, "------------------------------------------------------------------\n");
+ pcre_printint(re, outfile);
+ }
+
if (do_showinfo)
{
unsigned long int get_options, all_options;
int nameentrysize, namecount;
const uschar *nametable;
- if (do_debug)
- {
- fprintf(outfile, "------------------------------------------------------------------\n");
- _pcre_printint(re, outfile);
- }
-
new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options);
new_info(re, NULL, PCRE_INFO_SIZE, &size);
new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count);
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 (get_options == 0) fprintf(outfile, "No options\n");
- else fprintf(outfile, "Options:%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\n",
((get_options & PCRE_ANCHORED) != 0)? " anchored" : "",
((get_options & PCRE_CASELESS) != 0)? " caseless" : "",
((get_options & PCRE_EXTENDED) != 0)? " extended" : "",
((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "",
((get_options & PCRE_EXTRA) != 0)? " extra" : "",
((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "",
+ ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "",
((get_options & PCRE_UTF8) != 0)? " utf8" : "",
- ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "");
+ ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "",
+ ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : "");
+
+ switch (get_options & PCRE_NEWLINE_BITS)
+ {
+ case PCRE_NEWLINE_CR:
+ fprintf(outfile, "Forced newline sequence: CR\n");
+ break;
- if (((((real_pcre *)re)->options) & PCRE_ICHANGED) != 0)
- fprintf(outfile, "Case state changes\n");
+ case PCRE_NEWLINE_LF:
+ fprintf(outfile, "Forced newline sequence: LF\n");
+ break;
+
+ case PCRE_NEWLINE_CRLF:
+ fprintf(outfile, "Forced newline sequence: CRLF\n");
+ break;
+
+ case PCRE_NEWLINE_ANY:
+ fprintf(outfile, "Forced newline sequence: ANY\n");
+ break;
+
+ default:
+ break;
+ }
if (first_char == -1)
{
- fprintf(outfile, "First char at start or follows \\n\n");
+ fprintf(outfile, "First char at start or follows newline\n");
}
else if (first_char < 0)
{
int ch = first_char & 255;
const char *caseless = ((first_char & REQ_CASELESS) == 0)?
"" : " (caseless)";
- if (isprint(ch))
+ if (PRINTHEX(ch))
fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless);
else
fprintf(outfile, "First char = %d%s\n", ch, caseless);
int ch = need_char & 255;
const char *caseless = ((need_char & REQ_CASELESS) == 0)?
"" : " (caseless)";
- if (isprint(ch))
+ if (PRINTHEX(ch))
fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless);
else
fprintf(outfile, "Need char = %d%s\n", ch, caseless);
fprintf(outfile, "\n ");
c = 2;
}
- if (isprint(i) && i != ' ')
+ if (PRINTHEX(i) && i != ' ')
{
fprintf(outfile, "%c ", i);
c += 2;
strerror(errno));
}
else fprintf(outfile, "Study data written to %s\n", to_file);
+
}
}
fclose(f);
for (;;)
{
- unsigned char *q;
- unsigned char *bptr = dbuffer;
+ uschar *q;
+ uschar *bptr = dbuffer;
int *use_offsets = offsets;
int use_size_offsets = size_offsets;
int callout_data = 0;
options = 0;
+ *copynames = 0;
+ *getnames = 0;
+
+ copynamesptr = copynames;
+ getnamesptr = getnames;
+
pcre_callout = callout;
first_callout = 1;
callout_extra = 0;
callout_fail_id = -1;
show_malloc = 0;
- if (infile == stdin) printf("data> ");
- if (fgets((char *)buffer, BUFFER_SIZE, infile) == NULL)
+ if (extra != NULL) extra->flags &=
+ ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION);
+
+ len = 0;
+ for (;;)
{
- done = 1;
- goto CONTINUE;
+ if (infile == stdin) printf("data> ");
+ if (extend_inputline(infile, buffer + len) == NULL)
+ {
+ if (len > 0) break;
+ done = 1;
+ goto CONTINUE;
+ }
+ if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
+ len = (int)strlen((char *)buffer);
+ if (buffer[len-1] == '\n') break;
}
- if (infile != stdin) fprintf(outfile, "%s", (char *)buffer);
- len = (int)strlen((char *)buffer);
while (len > 0 && isspace(buffer[len-1])) len--;
buffer[len] = 0;
if (len == 0) break;
c -= '0';
while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9')
c = c * 8 + *p++ - '0';
+
+#if !defined NOUTF8
+ if (use_utf8 && c > 255)
+ {
+ unsigned char buff8[8];
+ int ii, utn;
+ utn = ord2utf8(c, buff8);
+ for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
+ c = buff8[ii]; /* Last byte */
+ }
+#endif
break;
case 'x':
{
unsigned char buff8[8];
int ii, utn;
- utn = _pcre_ord2utf8(c, buff8);
+ utn = ord2utf8(c, buff8);
for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii];
c = buff8[ii]; /* Last byte */
p = pt + 1;
}
else if (isalnum(*p))
{
- uschar name[256];
- uschar *npp = name;
+ uschar *npp = copynamesptr;
while (isalnum(*p)) *npp++ = *p++;
+ *npp++ = 0;
*npp = 0;
- n = pcre_get_stringnumber(re, (char *)name);
+ n = pcre_get_stringnumber(re, (char *)copynamesptr);
if (n < 0)
- fprintf(outfile, "no parentheses with name \"%s\"\n", name);
- else copystrings |= 1 << n;
+ fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr);
+ copynamesptr = npp;
}
else if (*p == '+')
{
}
else if (isalnum(*p))
{
- uschar name[256];
- uschar *npp = name;
+ uschar *npp = getnamesptr;
while (isalnum(*p)) *npp++ = *p++;
+ *npp++ = 0;
*npp = 0;
- n = pcre_get_stringnumber(re, (char *)name);
+ n = pcre_get_stringnumber(re, (char *)getnamesptr);
if (n < 0)
- fprintf(outfile, "no parentheses with name \"%s\"\n", name);
- else getstrings |= 1 << n;
+ fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr);
+ getnamesptr = npp;
}
continue;
options |= PCRE_PARTIAL;
continue;
+ case 'Q':
+ while(isdigit(*p)) n = n * 10 + *p++ - '0';
+ if (extra == NULL)
+ {
+ extra = (pcre_extra *)malloc(sizeof(pcre_extra));
+ extra->flags = 0;
+ }
+ extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
+ extra->match_limit_recursion = n;
+ continue;
+
+ case 'q':
+ while(isdigit(*p)) n = n * 10 + *p++ - '0';
+ if (extra == NULL)
+ {
+ extra = (pcre_extra *)malloc(sizeof(pcre_extra));
+ extra->flags = 0;
+ }
+ extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
+ extra->match_limit = n;
+ continue;
+
#if !defined NODFA
case 'R':
options |= PCRE_DFA_RESTART;
case '?':
options |= PCRE_NO_UTF8_CHECK;
continue;
+
+ case '<':
+ {
+ int x = check_newline(p, outfile);
+ if (x == 0) goto NEXT_DATA;
+ options |= x;
+ while (*p++ != '>');
+ }
+ continue;
}
*q++ = c;
}
if (rc != 0)
{
- (void)regerror(rc, &preg, (char *)buffer, BUFFER_SIZE);
+ (void)regerror(rc, &preg, (char *)buffer, buffer_size);
fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer);
}
+ else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE)
+ != 0)
+ {
+ fprintf(outfile, "Matched with REG_NOSUB\n");
+ }
else
{
size_t i;
for (;; gmatched++) /* Loop for /g or /G */
{
- if (timeit)
+ if (timeitm > 0)
{
register int i;
clock_t time_taken;
if (all_use_dfa || use_dfa)
{
int workspace[1000];
- for (i = 0; i < LOOPREPEAT; i++)
+ for (i = 0; i < timeitm; i++)
count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset,
options | g_notempty, use_offsets, use_size_offsets, workspace,
sizeof(workspace)/sizeof(int));
else
#endif
- for (i = 0; i < LOOPREPEAT; i++)
+ for (i = 0; i < timeitm; i++)
count = pcre_exec(re, extra, (char *)bptr, len,
start_offset, options | g_notempty, use_offsets, use_size_offsets);
time_taken = clock() - start_time;
- fprintf(outfile, "Execute time %.3f milliseconds\n",
- (((double)time_taken * 1000.0) / (double)LOOPREPEAT) /
+ fprintf(outfile, "Execute time %.4f milliseconds\n",
+ (((double)time_taken * 1000.0) / (double)timeitm) /
(double)CLOCKS_PER_SEC);
}
/* If find_match_limit is set, we want to do repeated matches with
- varying limits in order to find the minimum value. */
+ varying limits in order to find the minimum value for the match limit and
+ for the recursion limit. */
if (find_match_limit)
{
- int min = 0;
- int mid = 64;
- int max = -1;
-
if (extra == NULL)
{
extra = (pcre_extra *)malloc(sizeof(pcre_extra));
extra->flags = 0;
}
- extra->flags |= PCRE_EXTRA_MATCH_LIMIT;
- for (;;)
- {
- extra->match_limit = mid;
- count = pcre_exec(re, extra, (char *)bptr, len, start_offset,
- options | g_notempty, use_offsets, use_size_offsets);
- if (count == PCRE_ERROR_MATCHLIMIT)
- {
- /* fprintf(outfile, "Testing match limit = %d\n", mid); */
- min = mid;
- mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2;
- }
- else if (count >= 0 || count == PCRE_ERROR_NOMATCH ||
- count == PCRE_ERROR_PARTIAL)
- {
- if (mid == min + 1)
- {
- fprintf(outfile, "Minimum match limit = %d\n", mid);
- break;
- }
- /* fprintf(outfile, "Testing match limit = %d\n", mid); */
- max = mid;
- mid = (min + mid)/2;
- }
- else break; /* Some other error */
- }
+ (void)check_match_limit(re, extra, bptr, len, start_offset,
+ options|g_notempty, use_offsets, use_size_offsets,
+ PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit),
+ PCRE_ERROR_MATCHLIMIT, "match()");
- extra->flags &= ~PCRE_EXTRA_MATCH_LIMIT;
+ count = check_match_limit(re, extra, bptr, len, start_offset,
+ options|g_notempty, use_offsets, use_size_offsets,
+ PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion),
+ PCRE_ERROR_RECURSIONLIMIT, "match() recursion");
}
/* If callout_data is set, use the interface with additional data */
if (count >= 0)
{
- int i;
+ int i, maxcount;
+
+#if !defined NODFA
+ if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else
+#endif
+ maxcount = use_size_offsets/3;
+
+ /* This is a check against a lunatic return value. */
+
+ if (count > maxcount)
+ {
+ fprintf(outfile,
+ "** PCRE error: returned count %d is too big for offset size %d\n",
+ count, use_size_offsets);
+ count = use_size_offsets/3;
+ if (do_g || do_G)
+ {
+ fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G');
+ do_g = do_G = FALSE; /* Break g/G loop */
+ }
+ }
+
for (i = 0; i < count * 2; i += 2)
{
if (use_offsets[i] < 0)
{
if ((copystrings & (1 << i)) != 0)
{
- char copybuffer[16];
+ char copybuffer[256];
int rc = pcre_copy_substring((char *)bptr, use_offsets, count,
i, copybuffer, sizeof(copybuffer));
if (rc < 0)
}
}
+ for (copynamesptr = copynames;
+ *copynamesptr != 0;
+ copynamesptr += (int)strlen((char*)copynamesptr) + 1)
+ {
+ char copybuffer[256];
+ int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets,
+ count, (char *)copynamesptr, copybuffer, sizeof(copybuffer));
+ if (rc < 0)
+ fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc);
+ else
+ fprintf(outfile, " C %s (%d) %s\n", copybuffer, rc, copynamesptr);
+ }
+
for (i = 0; i < 32; i++)
{
if ((getstrings & (1 << i)) != 0)
else
{
fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc);
- /* free((void *)substring); */
pcre_free_substring(substring);
}
}
}
+ for (getnamesptr = getnames;
+ *getnamesptr != 0;
+ getnamesptr += (int)strlen((char*)getnamesptr) + 1)
+ {
+ const char *substring;
+ int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets,
+ count, (char *)getnamesptr, &substring);
+ if (rc < 0)
+ fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc);
+ else
+ {
+ fprintf(outfile, " G %s (%d) %s\n", substring, rc, getnamesptr);
+ pcre_free_substring(substring);
+ }
+ }
+
if (getlist)
{
const char **stringlist;
len -= use_offsets[1];
}
} /* End of loop for /g and /G */
+
+ NEXT_DATA: continue;
} /* End of loop for data lines */
CONTINUE:
{
new_free((void *)tables);
setlocale(LC_CTYPE, "C");
+ locale_set = 0;
}
}