On Mon, Mar 20, 2017 at 9:48 AM, Florian Weimer <fweimer@redhat.com> wrote:
On 03/20/2017 02:40 PM, Carlos O'Donell wrote:
On Mon, Mar 20, 2017 at 8:13 AM, Florian Weimer <fweimer@redhat.com>
wrote:
On 03/18/2017 03:13 AM, Patrick wrote:
Could anyone tell me what I should do next? Is it hard to create a patch
for this?
It is very hard. We no longer have a working non-Linux port of glibc.
All
non-Linux ports (Hurd, FreeBSD, NaCl, maybe there are more) are either
out-of-tree or impossibly difficult to compile.
My theory is that this is a reflection of the interest and resources
available to the porters rather than anything inherent in the
complexity or drastic differences between the OSs.
I meant that today, there is only Linux as a port which sees regular
compile-time testing at least. We can't really write generic code which
supports non-Linux platforms if there isn't a single non-Linux platform
which we can compile as a target. We try to catch potential issues during
code review, but that's not always going to prevent Linux-isms from seeping
in.
This means that ironing out all the hidden Linux-isms in the code base is a
lot of work.
That is very true. We really need a complete GNU/Linux-hosted
toolchain to do cross-compile tests with the
scripts/build-many-glibcs.py test script.
Cheers,
Carlos.