[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: dfa.h / dfa.c diff versus gawk attached

From: Tony Abou-Assaleh
Subject: Re: dfa.h / dfa.c diff versus gawk attached
Date: Tue, 23 Oct 2007 02:22:32 -0300 (ADT)

> > Is there a way to pass grep a pattern with an embedded new lines from the
> > command line? Or do I really have to write some code that uses the dfa
> > directly to test this?
> You can always use
>       grep 'some
>       text' file1 ...

It seems that grep treats new lines as pattern separators. The above is
currently equivalent to:

        grep -e 'some' -e 'text' file1 ...

I haven't looked yet to see where the interpretation of newlines is done.
Further, man grep for -f states:

        "Obtain patterns from FILE, one per line"

We will need to add a meta string representing new lines if we don't want
to break backward compatibility. Another option is to add something
similar to -z but for patterns, but then patterns won't be able to embed
null characters for binary matching.

> I can, MAYBE, split it out into 2 patches: restore-newline-matching
> and cumultaive-bug-fixes, but I don't have the energy for any more than
> that.

Sounds good. Thanks.

> I thought I added a comment. If not, the code is clear enough. A pattern
> of the form    ab{0}c   is treated by dfa as   ab{1}c   and incorrectly
> does not match "ac".  Regex gets this right.  This is the only case I've
> found so far where dfa is broken, but if I find others, at least I now
> have a way to signal the caller.

Ah, I see. I wasn't sure if it was related to this bug or something else.
The good news is that now we can have some test cases :O)



Tony Abou-Assaleh
Email:    address@hidden
Web site: http://tony.abou-assaleh.net
----------------------[THE END]----------------------

reply via email to

[Prev in Thread] Current Thread [Next in Thread]