Re: grep -i bug

[Chris F.A. Johnson]

On Tue, 15 Feb 2005, Karsten Johansson wrote:

> Greetings,
>
> I have found what looks like a bug in grep. I have not been able to test it on
> other systems, but for the record, I am using SuSE's Novell Linux. I haven't
> noticed this bug before, but then again, I hadn't tried to grep through a 40
> Gigabyte file before.
>
> Steps to see the bug:
>
>     grep -i someword some40Gigabitefile.txt > results.txt
>
> looking at results, we could for example see:
>
>     40047878321- someword
>
> Now do this:
>     xxd -s 40047878321 | less
>
> This /should/ take us right to the offset where the word appears. For the most
> part it does. But once the word is found really deep into the file - ie: the
> byte offset gets huge - those offsets have nothing to do with reality.
>
> Let me know if you have already discovered this, if it is fixed, or what I'm
> doing wrong to make this happen.

    What size integer does xxd use? Could it be an overflow error in
    xxd?

    Have you tried that byte offset with other utilities?

--
        Chris F.A. Johnson                      http://cfaj.freeshell.org
        =================================================================
                Everything in moderation -- including moderation