This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: RFC: should we use -Werror? (& sample patch to do it)
- To: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- Subject: Re: RFC: should we use -Werror? (& sample patch to do it)
- From: Geoff Keating <geoffk at geoffk dot org>
- Date: 05 Sep 2001 07:15:11 -0700
- CC: gcc at gcc dot gnu dot org
- References: <200109050332.XAA14648@caip.rutgers.edu>
"Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> writes:
> I got warning levels down to around 30 on solaris2.7 and 20 on irix6.
> So I thought it might be worth a try to get -Werror installed in gcc.
>
> There are some problems. We don't want to -Werror stage1 when using
> e.g. 2.95 to bootstrap because older GCCs gives many false positives
> that have since been corrected. Using 2.95 on stage1 yields over 350
> warnings on the same solaris2.7 that gets just 30 in all of stage3
> (and stage3 builds all languges, not just C!)
>
> Thus IMHO, we should only enable -Werror in stage2 or later. However
> this means if you don't do a full bootstrap (e.g. cross-compiles or
> building for embedded targets) then -Werror isn't in effect. This
> also means the automatic tester to powerpc won't detect regressions.
That's OK. The automatic tester also does an x86-linux bootstrap.
--
- Geoffrey Keating <geoffk@geoffk.org>