Revision: 662
http://www.exim.org/viewvc/pcre2?view=rev&revision=662
Author: ph10
Date: 2017-02-10 17:39:29 +0000 (Fri, 10 Feb 2017)
Log Message:
-----------
Fix previously broken fix for pcre2grep with -Mo matching strings that cross
line boundaries.
Modified Paths:
--------------
code/trunk/ChangeLog
code/trunk/RunGrepTest
code/trunk/src/pcre2grep.c
code/trunk/testdata/grepoutput
Modified: code/trunk/ChangeLog
===================================================================
--- code/trunk/ChangeLog 2017-02-10 16:42:14 UTC (rev 661)
+++ code/trunk/ChangeLog 2017-02-10 17:39:29 UTC (rev 662)
@@ -352,7 +352,12 @@
complicated that this optimization probably isn't worth it.) This fixes
oss-fuzz issue 557.
+55. Issue 32 for 10.22 below was not correctly fixed. If pcre2grep in multiline
+mode with --only-matching matched several lines, it restarted scanning at the
+next line instead of moving on to the end of the matched string, which can be
+several lines after the start.
+
Version 10.22 29-July-2016
--------------------------
Modified: code/trunk/RunGrepTest
===================================================================
--- code/trunk/RunGrepTest 2017-02-10 16:42:14 UTC (rev 661)
+++ code/trunk/RunGrepTest 2017-02-10 17:39:29 UTC (rev 662)
@@ -593,6 +593,11 @@
(cd $srcdir; $valgrind $vjs $pcre2grep -tL 'the' testdata/grepinput*) >>testtrygrep
echo "RC=$?" >>testtrygrep
+echo "---------------------------- Test 119 -----------------------------" >>testtrygrep
+printf "123\n456\n789\n---abc\ndef\nxyz\n---\n" >testNinputgrep
+(cd $srcdir; $valgrind $vjs $pcre2grep -Mo '(\n|[^-])*---' testNinputgrep) >>testtrygrep
+echo "RC=$?" >>testtrygrep
+
# Now compare the results.
$cf $srcdir/testdata/grepoutput testtrygrep
Modified: code/trunk/src/pcre2grep.c
===================================================================
--- code/trunk/src/pcre2grep.c 2017-02-10 16:42:14 UTC (rev 661)
+++ code/trunk/src/pcre2grep.c 2017-02-10 17:39:29 UTC (rev 662)
@@ -2285,11 +2285,6 @@
if (line_buffered) fflush(stdout);
rc = 0; /* Had some success */
- /* If the current match ended past the end of the line (only possible
- in multiline mode), we are done with this line. */
-
- if (offsets[1] > linelength) goto END_ONE_MATCH;
-
/* If the pattern contained a lookbehind that included \K, it is
possible that the end of the match might be at or before the actual
starting offset we have just used. In this case, start one character
@@ -2301,9 +2296,23 @@
{
if (startoffset >= length) goto END_ONE_MATCH; /* Were at end */
startoffset = oldstartoffset + 1;
- if (utf)
- while ((matchptr[startoffset] & 0xc0) == 0x80) startoffset++;
+ if (utf) while ((matchptr[startoffset] & 0xc0) == 0x80) startoffset++;
}
+
+ /* If the current match ended past the end of the line (only possible
+ in multiline mode), we must move on to the line in which it did end
+ before searching for more matches. */
+
+ while (startoffset > linelength)
+ {
+ matchptr = ptr += linelength + endlinelength;
+ filepos += (int)(linelength + endlinelength);
+ linenumber++;
+ startoffset -= (int)(linelength + endlinelength);
+ t = end_of_line(ptr, endptr, &endlinelength);
+ linelength = t - ptr - endlinelength;
+ }
+
goto ONLY_MATCHING_RESTART;
}
}
Modified: code/trunk/testdata/grepoutput
===================================================================
(Binary files differ)
