Package: pcregrep Version: 6.4-1.1 Severity: normal When pcregrep is outputting the input lines that matched the regexp, it truncates each output line at the first null character. AFAICT, the bug affects only the output: the regexp engine correctly finds characters that follow a null character in an input line, and it is even possible to search for the null character itself.
[EMAIL PROTECTED]:~$ printf "pot\0kettle\nblack\n" | hd
00000000 70 6f 74 00 6b 65 74 74 6c 65 0a 62 6c 61 63 6b |pot.kettle.black|
00000010 0a |.|
00000011
[EMAIL PROTECTED]:~$ printf "pot\0kettle\nblack\n" | pcregrep k | hd
00000000 70 6f 74 0a 62 6c 61 63 6b 0a |pot.black.|
0000000a
[EMAIL PROTECTED]:~$ printf "pot\0kettle\nblack\n" | pcregrep '\0' | hd
00000000 70 6f 74 0a |pot.|
00000004
[EMAIL PROTECTED]:~$
The bug may be in this line in the pcregrep function in pcregrep.c:
fprintf(stdout, "%.*s\n", linelength, ptr);
Changing this to fwrite(ptr, 1, linelength, stdout) might fix the
bug, but I haven't tested it. I believe this change would not
affect the handling of multibyte characters.
There may also be similar bugs in the output of neighbouring lines
(the --after-context=number and --before-context=number options).
-- System Information:
Debian Release: testing/unstable
APT prefers unstable
APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.12-1-k7
Locale: LANG=fi_FI.UTF-8, LC_CTYPE=fi_FI.UTF-8 (charmap=UTF-8)
Versions of packages pcregrep depends on:
ii libc6 2.3.5-6 GNU C Library: Shared libraries an
ii libpcre3 6.4-1.1 Perl 5 Compatible Regular Expressi
pcregrep recommends no packages.
-- no debconf information
pgpPXJ1AjZ7yP.pgp
Description: PGP signature
