Revision: 1155
http://www.exim.org/viewvc/pcre2?view=rev&revision=1155
Author: ph10
Date: 2019-08-03 09:30:40 +0100 (Sat, 03 Aug 2019)
Log Message:
-----------
Fix incorrect computation of group length when one branch exceeded 65535.
Modified Paths:
--------------
code/trunk/ChangeLog
code/trunk/src/pcre2_study.c
code/trunk/testdata/testinput2
code/trunk/testdata/testoutput2
Modified: code/trunk/ChangeLog
===================================================================
--- code/trunk/ChangeLog 2019-08-01 16:59:50 UTC (rev 1154)
+++ code/trunk/ChangeLog 2019-08-03 08:30:40 UTC (rev 1155)
@@ -124,7 +124,12 @@
26. Make pcre2test -C show if libreadline or libedit is supported.
+28. If the length of one branch of a group exceeded 65535 (the maximum value
+that is remembered as a minimum length), the whole group's length was
+incorrectly recorded as 65535, leading to incorrect "no match" when start-up
+optimizations were in force.
+
Version 10.33 16-April-2019
---------------------------
Modified: code/trunk/src/pcre2_study.c
===================================================================
--- code/trunk/src/pcre2_study.c 2019-08-01 16:59:50 UTC (rev 1154)
+++ code/trunk/src/pcre2_study.c 2019-08-03 08:30:40 UTC (rev 1155)
@@ -105,6 +105,7 @@
int *backref_cache)
{
int length = -1;
+int branchlength = 0;
int prev_cap_recno = -1;
int prev_cap_d = 0;
int prev_recurse_recno = -1;
@@ -112,9 +113,9 @@
uint32_t once_fudge = 0;
BOOL had_recurse = FALSE;
BOOL dupcapused = (re->flags & PCRE2_DUPCAPUSED) != 0;
+PCRE2_SPTR nextbranch = code + GET(code, 1);
+PCRE2_UCHAR *cc = (PCRE2_UCHAR *)code + 1 + LINK_SIZE;
recurse_check this_recurse;
-int branchlength = 0;
-PCRE2_UCHAR *cc = (PCRE2_UCHAR *)code + 1 + LINK_SIZE;
/* If this is a "could be empty" group, its minimum length is 0. */
@@ -130,16 +131,20 @@
/* Scan along the opcodes for this branch. If we get to the end of the branch,
check the length against that of the other branches. If the accumulated length
-passes 16-bits, stop. */
+passes 16-bits, reset to that value and skip the rest of the branch. */
for (;;)
{
int d, min, recno;
- PCRE2_UCHAR *cs, *ce;
- PCRE2_UCHAR op = *cc;
+ PCRE2_UCHAR op, *cs, *ce;
- if (branchlength >= UINT16_MAX) return UINT16_MAX;
+ if (branchlength >= UINT16_MAX)
+ {
+ branchlength = UINT16_MAX;
+ cc = (PCRE2_UCHAR *)nextbranch;
+ }
+ op = *cc;
switch (op)
{
case OP_COND:
@@ -229,6 +234,7 @@
if (length < 0 || (!had_recurse && branchlength < length))
length = branchlength;
if (op != OP_ALT) return length;
+ nextbranch = cc + GET(cc, 1);
cc += 1 + LINK_SIZE;
branchlength = 0;
had_recurse = FALSE;
@@ -241,7 +247,7 @@
case OP_ASSERTBACK:
case OP_ASSERTBACK_NOT:
case OP_ASSERT_NA:
- case OP_ASSERTBACK_NA:
+ case OP_ASSERTBACK_NA:
do cc += GET(cc, 1); while (*cc == OP_ALT);
/* Fall through */
@@ -1091,7 +1097,7 @@
case OP_ONCE:
case OP_SCRIPT_RUN:
case OP_ASSERT:
- case OP_ASSERT_NA:
+ case OP_ASSERT_NA:
rc = set_start_bits(re, tcode, utf);
if (rc == SSB_FAIL || rc == SSB_UNKNOWN) return rc;
if (rc == SSB_DONE) try_next = FALSE; else
@@ -1134,7 +1140,7 @@
case OP_ASSERT_NOT:
case OP_ASSERTBACK:
case OP_ASSERTBACK_NOT:
- case OP_ASSERTBACK_NA:
+ case OP_ASSERTBACK_NA:
do tcode += GET(tcode, 1); while (*tcode == OP_ALT);
tcode += 1 + LINK_SIZE;
break;
@@ -1584,9 +1590,9 @@
/* This function is handed a compiled expression that it must study to produce
information that will speed up the matching.
-Argument:
+Argument:
re points to the compiled expression
-
+
Returns: 0 normally; non-zero should never normally occur
1 unknown opcode in set_start_bits
2 missing capturing bracket
Modified: code/trunk/testdata/testinput2
===================================================================
--- code/trunk/testdata/testinput2 2019-08-01 16:59:50 UTC (rev 1154)
+++ code/trunk/testdata/testinput2 2019-08-03 08:30:40 UTC (rev 1155)
@@ -5740,4 +5740,12 @@
/c*+/
ab\=ph,offset=2
+/\A\s*(a|(?:[^`]{28500}){4})/I
+ a
+
+/\A\s*((?:[^`]{28500}){4})/I
+
+/\A\s*((?:[^`]{28500}){4}|a)/I
+ a
+
# End of testinput2
Modified: code/trunk/testdata/testoutput2
===================================================================
--- code/trunk/testdata/testoutput2 2019-08-01 16:59:50 UTC (rev 1154)
+++ code/trunk/testdata/testoutput2 2019-08-03 08:30:40 UTC (rev 1155)
@@ -17267,6 +17267,33 @@
ab\=ph,offset=2
Partial match:
+/\A\s*(a|(?:[^`]{28500}){4})/I
+Capture group count = 1
+Max lookbehind = 1
+Compile options: <none>
+Overall options: anchored
+Subject length lower bound = 1
+ a
+ 0: a
+ 1: a
+
+/\A\s*((?:[^`]{28500}){4})/I
+Capture group count = 1
+Max lookbehind = 1
+Compile options: <none>
+Overall options: anchored
+Subject length lower bound = 65535
+
+/\A\s*((?:[^`]{28500}){4}|a)/I
+Capture group count = 1
+Max lookbehind = 1
+Compile options: <none>
+Overall options: anchored
+Subject length lower bound = 1
+ a
+ 0: a
+ 1: a
+
# End of testinput2
Error -70: PCRE2_ERROR_BADDATA (unknown error number)
Error -62: bad serialized data
