GNU bug report logs - #25027
Grep 2.25 misses lines in semi-large file

Previous Next

To reply to this bug, email your comments to 25027 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Mattias Jiderhamn <mattias.jiderhamn <at> lexmark.com>
To: bug-grep <at> gnu.org
Subject: Grep 2.25 misses lines in semi-large file
Date: Fri, 25 Nov 2016 14:25:46 +0100

[Message part 1 (text/plain, inline)]

Hi. I'm using grep 2.25 in cygwin on a 17 MB log file. I happened to 
notice that when I did "grep Migrate" on the file some lines were 
missing. If however I did some other filtering first, i.e. "grep foo | 
grep Migrate" (where foo would match those lines), they are included.

I cannot reproduce this issue with grep 2.5.4 on Linux.

I tried minimizing the log file by removing lines that I did not expect 
in the output, but when removing lines before the expected line the 
problem dissappeared, so somewhere in the first 17 MB (of my 50 MB file) 
there is something triggering the bug.

Due to the nature of the logged data, I'm not willing to make the file 
publically available in a bug report, but I'm willing to transfer it to 
one of the maintainers, or by some other means help out in debugging. 
Please suggest how to proceed.
-- 

*Mattias Jiderhamn*
Architect & Development Lead
Invoice Portal & Exder

*M* +46 735 16 34 13
mattias.jiderhamn <at> lexmark.com

<http://www.lexmark.com>
www.lexmark.com

[Message part 2 (text/html, inline)]

[lexmark-logo-email-signature.png (image/png, inline)]

Message #8 received at 25027 <at> debbugs.gnu.org (full text, mbox):

From: Jim Meyering <jim <at> meyering.net>
To: Mattias Jiderhamn <mattias.jiderhamn <at> lexmark.com>
Cc: 25027 <at> debbugs.gnu.org
Subject: Re: bug#25027: Grep 2.25 misses lines in semi-large file
Date: Fri, 25 Nov 2016 10:13:27 -0800

On Fri, Nov 25, 2016 at 5:25 AM, Mattias Jiderhamn
<mattias.jiderhamn <at> lexmark.com> wrote:
> Hi. I'm using grep 2.25 in cygwin on a 17 MB log file. I happened to notice
> that when I did "grep Migrate" on the file some lines were missing. If
> however I did some other filtering first, i.e. "grep foo | grep Migrate"
> (where foo would match those lines), they are included.
>
> I cannot reproduce this issue with grep 2.5.4 on Linux.
>
> I tried minimizing the log file by removing lines that I did not expect in
> the output, but when removing lines before the expected line the problem
> dissappeared, so somewhere in the first 17 MB (of my 50 MB file) there is
> something triggering the bug.
>
> Due to the nature of the logged data, I'm not willing to make the file
> publically available in a bug report, but I'm willing to transfer it to one
> of the maintainers, or by some other means help out in debugging. Please
> suggest how to proceed.

Thank you for the report.

It may be useful to know what locale you are using in those tests.
I.e., what does "locale" print for you? If you're not already using
the "C" locale, can you reproduce the problem with "LC_ALL=C grep
Migrate" ?

If that doesn't help, I'd be happy to investigate. You are welcome to
send me that log file, preferably well-compressed.

Message #11 received at 25027 <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Jim Meyering <jim <at> meyering.net>,
 Mattias Jiderhamn <mattias.jiderhamn <at> lexmark.com>
Cc: 25027 <at> debbugs.gnu.org
Subject: Re: bug#25027: Grep 2.25 misses lines in semi-large file
Date: Fri, 25 Nov 2016 10:17:06 -0800

Another possibility is to try to sanitize the data. If you're using the C 
locale, for example, replace all  characters in the set A-Za-z0-9_ (other than 
that crucial "Migrate") with 'A', replace all other non-space characters with 
'.', etc. If you're using UTF-8, make sure any replacements do not change the 
number of bytes in the character. If you can reproduce the problem with the 
sanitized data, compress it and email it as an attachment.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.