This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Add hidden visibility to internal function prototypes
On Tue, 22 Aug 2017, Florian Weimer wrote:
> > As far as I know, linking those tests statically would be OK (it would not
> > significantly affect what they are testing). That's atext-exp,
> > atest-sincos, atest-exp2 that are using internal GMP objects from libc.a;
> > I don't know if other tests, in math/ or elsewhere, also link against
> > particular libc.a objects.
>
> I think the attached approach will work for the math subdirectory.
> Tested on aarch64 and ppc64le.
>
> I see additional aarch64 failures, though (outside math).
To be clear: it works for math/ on top of having reapplied HJ's patch, but
there are still similar failures in other directories (on aarch64) in the
presence of HJ's patch?
What are those other aarch64 failures? Maybe we should look at the full
set of failures in all directories with the visibility changes in order to
figure out what approaches are good for fixing such problems.
In principle it should make sense for all symbols in the static libraries
to be hidden (maybe build them with -fvisibility=hidden?), unless there's
a specific reason some given symbol needs to be non-hidden for normal use
of the installed libraries, but first we need to sort out any testsuite
breakage that causes.
--
Joseph S. Myers
joseph@codesourcery.com