Re: [PATCH 2/2] Add missing argz_* functions from glibc

[Jim Meyering]

Eric Blake <address@hidden> wrote:
> Jim Meyering <jim <at> meyering.net> writes:
>> > Yeah, it seems silly to test for all of them; wouldn't it make sense to
>> > check for just one of them (the last one added) ?
>>
>> argz_replace is the one with the most recent first log entry:
>
> For glibc.  On cygwin, it looks like all the argz_ functions were added at
> once, on 2002-06-14.  But does this properly test for bugs in the argz_
> implementation?  For example, cygwin argz_add_sep failed to handle empty 
> string
> arguments as recently as 2007-05-29.  Libtool has already faced hurdles
> detecting cygwin brokenness in the subset of argz_ that libltdl uses; and now
> that we are expanding the set of functions, we should be sure that the .m4 
> file
> catches all of cygwin's previous bugs, and not just the ones tickled by 
> libltdl.

That sounds like a nice addition, for cygwin users.
Maybe it's enough just to check for just the latest or a few key bugs?
You're in the best position to answer that one.

In the mean time, I'm going to check in the new generated files in an
hour or two.  So if anyone has an objection, speak up soon.
Lower priority: I'll look into automating the procedure
to generate these files.