bug-grep
[Top][All Lists]
Advanced

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

Re: FYI, [PATCH] use gnulib's regex code


From: Jim Meyering
Subject: Re: FYI, [PATCH] use gnulib's regex code
Date: Thu, 04 Feb 2010 14:52:35 +0100

Paolo Bonzini wrote:
> On 02/04/2010 08:55 AM, Jim Meyering wrote:
>> +        $(PERL) -w$$T_ -I$(srcdir) -MCoreutils                              
>> \
>> +          -M"CuTmpdir qw($$f)" -- "$$1";    \
>
> cut'n'paste?

No, it was deliberate, albeit perhaps getting ahead of myself.

That code is currently dead.
I find the CuTmpdir.pm + Coreutils.pm framework to be more robust
and maintainable than awk+shell-based ones for some classes of tests,
perhaps like the spencer ones, so opted to leave that in for now.
Then, if I add new tests like that, it's less work to use that framework.

> You know that I don't like the signedness patches to gnulib's reg*.c
> but this is not the place to discuss them anyway (and if I can

One of them is now gone, since I found a better way to avoid
a problematic warning in regcomp.c: the patch no longer
requires

#include "intprops.h"
#include "verify.h"

for this use:

  verify (! TYPE_SIGNED (Idx));
  if (end != REG_MISSING)

Instead, the regcomp.c now in gnulib does this:

#define TYPE_SIGNED(t) (! ((t) 0 < (t) -1))
  if (TYPE_SIGNED (Idx) || end != REG_MISSING)

> convince you to change the definition of Idx, grep won't break because
> the submodule will still point to a gnulib that can receive the
> patch).  So, please go ahead!

Thanks for looking.




reply via email to

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