Bug 112869

Summary:	grep -iE segfaults using \W unless LANG=C
Product:	[Fedora] Fedora	Reporter:	Davide Bolcioni <davide_bolcioni>
Component:	glibc	Assignee:	Jakub Jelinek <jakub>
Status:	CLOSED ERRATA	QA Contact:	Brian Brock <bbrock>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	1	CC:	twaugh
Target Milestone:	---
Target Release:	---
Hardware:	athlon
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2004-01-14 15:49:44 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Davide Bolcioni 2004-01-04 22:09:03 UTC

Description of problem: Attempting to use grep -iE with \W in the 
pattern segfaults with LANG=it_IT, but not with LANG=C.


Version-Release number of selected component (if applicable): 2.5.1-17.4

How reproducible: Always


Steps to Reproduce:
1. echo '<form>' | LANG=it_IT grep -iE '\Wform\W' segfaults
2.
3.
  
Actual results: segfaults.


Expected results: no segfault and outputs <form>.


Additional info:  echo '<form>' | LANG=C grep -iE '\Wform\W' does not
segfault and works as expected.

Comment 1 Davide Bolcioni 2004-01-04 22:10:18 UTC

If -i is omitted, works as expected.

Comment 2 Tim Waugh 2004-01-05 12:15:13 UTC

This seems to be a glibc bug, as shown by:

#include <locale.h>
#include <regex.h>
#include <ctype.h>
int main (void)
{
  struct re_pattern_buffer re = { 0 };
  unsigned char trans[256];
  int i;
  setlocale (LC_ALL, "en_GB");
  for (i = 0; i < 256; ++i) trans[i] = tolower (i);
  re.translate = trans;
  re_set_syntax (RE_SYNTAX_POSIX_EGREP);
  re_compile_pattern ("\\W", 2, &re);
}

However, I have built grep-2.5.1-24 with a work-around to avoid this bug.

Changing component to glibc and reassigning.

Comment 3 Jakub Jelinek 2004-01-14 15:49:44 UTC

Should be fixed in glibc-2.3.2-101.4.