This is the mail archive of the
libc-hacker@sourceware.cygnus.com
mailing list for the glibc project.
Re: <signal.h> broken with _POSIX_SOURCE=1 defined
- To: Andreas Schwab <schwab@issan.informatik.uni-dortmund.de>
- Subject: Re: <signal.h> broken with _POSIX_SOURCE=1 defined
- From: Zack Weinberg <zack@rabi.columbia.edu>
- Date: Tue, 03 Nov 1998 19:36:04 -0500
- cc: libc-hacker@cygnus.com
On 03 Nov 1998 16:14:22 +0100, Andreas Schwab wrote:
>Zack Weinberg <zack@rabi.columbia.edu> writes:
>
>|> On Tue, 03 Nov 1998 08:08:07 -0500, Zack Weinberg wrote:
>|> >
>|> >As an experiment I wrote a simple test which just includes every
>|> >single public header we provide with _GNU_SOURCE defined.
>|>
>|> Followup: Doing the same thing with _POSIX_SOURCE instead produces
>|> some spurious errors. They come from the internal prototypes in the
>|> wrapper headers.
>
>The library source largely depends on being compiled with _GNU_SOURCE.
>You can only get correct results by testing against installed headers.
I know and I consider it broken. This is what Roland was worried
about back in August when libc's view of some headers was made
different from the normal-program view. A test case has to be treated
as if it were a normal program.
>From the end-user point of view, you don't want to be told "you can't
run these tests of libc until after you've installed it and clobbered
your old libc".
I know how to fix the problem, but the code changes would be so
extensive I don't think Uli wants them in 2.1.
zw