This is the mail archive of the gcc-patches@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: [3.4 PATCH] Fix powerpc*-*-linux* bootstrap (PR target/19019)


On Thu, Feb 24, 2005 at 10:01:56AM +1030, Alan Modra wrote:
> Looks OK to me too.  I was developing a competing patch when I saw
> Jakub's post last night, but was stuck in shell quoting hell.  I think
> my patch is more elegant (I'm biased, of course!), but Jakub's patch is
> simpler in the sense that he doesn't try to use much shell magic.  It's
> probably safer to go with Jakub's patch for 3.4, and of course this
> patch isn't needed on mainline.  Anyway, just in case some problem
> arises with Jakub's patch, here's mine.

Well, on mainline is something needed too, see
http://gcc.gnu.org/ml/gcc-patches/2005-02/msg01418.html
Although GCC builds there, it is a bad idea to have versioned but hidden
symbols put into other binaries/shared libraries.
But IMHO passing -DSHARED to all *_s.o objects is desirable if not for
anything else then for consistency (it is already used e.g. for ia64 .symver
stuff).  And once the mklibgcc.in change is in, the ppc64 fix is a
one-liner.

FYI, I have bootstrapped/regtested both the 4.0 and 3.4 patch on 7 linux
arches (inclusing ppc and ppc64) each with no regressions.

	Jakub


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