This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Question on -Werror usage in Makefiles...


"Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> writes:

>  > Looking at the commandline, I wonder why we do not pass -Werror to
>  > this?  I'll send a patch for the actual warning later but would
>  > welcome if somebody fixes the Makefiles to actually use -Werror
>  > here.  Or is this on purpose?
>  > Andreas
>
> Yes, at the time I added -Werror I did that on purpose.  There was a
> lot of resistance to adding that flag and I felt I was personally
> being asked to fix any and all errors that resulted.  So I only
> activated it on "host" and "build" files, because those are easier to
> build in a cross compiler and I could preemptively fix a lot of the
> problems myself.  To trigger "target" file warnings, i.e. to build
> libgcc.a in a cross-compiler, you also have to build binutils and have
> the appropriate headers for each target.  I didn't have the time to do
> that for all targets (or even just one on each cpu) and preemptively
> fix those warnings.  So I took things as far as I could.

Thanks for the background.

> Since then there seems to be more acceptance of -Werror, and it may be
> that we can count on the community at large to fix their own warnings
> on their respective targets.  So there's no reason we can't use
> -Werror on target files too.  It's just a matter of someone
> volunteering to champion it and fix some of the most common problems.
> (I don't have the time myself to be point person.)  There will
> undoubtedly be some naysayers, but I think the feasibility of using
> -Werror so far has been self-evident.
>
> If you are interested I would encourage you to proceed.  I'll pitch in
> as time permits and back you up if you decide to give it a go.

Do you have a patch?  In that case I would suggest that some of us
apply it locally and try to kill those warnings that remain - and
let's then decide whether it's ok to add it per default.

Andreas
-- 
 Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj
  SuSE Linux AG, Deutschherrnstr. 15-19, 90429 Nürnberg, Germany
   GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126

Attachment: pgp00000.pgp
Description: PGP signature


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]