On Mon, 22 Feb 2016 12:55:56 +0100
Marek Polacek <polacek(a)redhat.com> wrote:
In the past few days and weeks we did a mass rebuild of Fedora
rawhide packages in mock with GCC 6 (and corresponding libtool) and
for those packages that failed also rebuilt the same package with
gcc-5.3.1-2.fc23.x86_64 to quickly remove from the list packages that
fail for non-GCC related reasons.
There were 17741 packages overall (last year we had 16230 packages).
16281 packages built fine, 883 packages failed to build with both
gcc-6 and gcc-5 (ignored for this analysis, unlikely to be GCC 6
related). This left us with 577 packages that had to be analyzed.
That is a lot -- last year we only had to examine 236 packages. So
that's why it's taken so long.
As usually, there will be a "porting to" document to ease the
transition to the new GCC. We already have
, even though this document
is still somewhat in flux.
The biggest change hands-down this year is the change of the default
standard for g++ from -std=gnu++98 to -std=gnu++14; this has caused
considerable churn, as you might have noticed on this mailing list.
Unfortunately, many packages weren't prepared to handle C++11.
Changes in libstdc++ often revealed very poor programming practices.
Before I describe the results in more detail, I'd like to thank Jon
Wakely and Jakub Jelinek for their indispensable help.
Any mistakes, omissions, or mis-categorizations are solely mine.
I see an inconsistency between s390(x) and other arches with "stdarg"
handling, build on s390(x) of a C++ lib fails with
error: first argument to ‘va_arg’ not of type ‘va_list’
Please see https://bugzilla.redhat.com/show_bug.cgi?id=1314545
details and a reproducer.