bug-glibc
[Top][All Lists]
Advanced

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

Re: 2.2.4, gsed != sed


From: Andreas Jaeger
Subject: Re: 2.2.4, gsed != sed
Date: Sat, 29 Dec 2001 17:35:38 +0100
User-agent: Gnus/5.090004 (Oort Gnus v0.04) XEmacs/21.4 (Artificial Intelligence, i386-suse-linux)

Ulrich Drepper <address@hidden> writes:

> Andreas Jaeger <address@hidden> writes:
>
>> configure.in has:
>> 
>> AC_CHECK_PROG_VER(SED, gsed sed, --version,
>>   [GNU sed version \([0-9]*\.[0-9.]*\)],
>>   [3.0[2-9]*|3.[1-9]*|[4-9]*],
>>   SED=: aux_missing="$aux_missing sed")
>
> I don't know why there is there is the test for gsed (probably copied
> from somewhere else) but the test alone is of course useful without
> using SED.  And the point stands: non-GNU seds would never be tested.

I agree, the test is usefull.  But I propose to remove the "gsed"
since gsed (and neither SED) is nowhere used and we use sed
everywhere.

> I.e., every release has the good chance of not working on any system
> where sed is not GNU sed despite this change.  Appearances of sed in
> Makefiles might be more easily detectable (and can be handled easily).

I agree.

> But sed is used in all kind of scripts.  Getting the configuration
> there (the equivalent of the SED variable) requires major changes.

That's what I also thought.

> And fact is that for all the time the code exists and is used this
> never was the problem.  So why bother?

Ok, so let's remove the "gsed" from configure.in...

Andreas
-- 
 Andreas Jaeger
  SuSE Labs address@hidden
   private address@hidden
    http://www.suse.de/~aj



reply via email to

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