Re: [lmi] Testing with clang (was: More clang [[noreturn]] changes)

[Vadim Zeitlin]

On Wed, 19 Apr 2017 23:22:16 +0000 Greg Chicares <address@hidden> wrote:

GC> On 2017-04-14 17:13, Vadim Zeitlin wrote:
GC> > 
GC> >  Sorry for not submitting both of them together with the other changes
GC> > before, but it turns out there are more instances of the same 
"[[noreturn]]
GC> > used in function definition and not declaration" problem, so here are two
GC> > more patches fixing them and allowing me to fully build lmi with clang.
GC> 
GC> Both committed and pushed.

 Perfect, thanks!

GC> Maybe we should skip gcc-5 and upgrade to gcc-6.2.1 because that's what
GC> debian-9 offers.

 Yes, this would make sense. Although gcc 7 might be expected to find its
way in backports (i.e. be easily installable) too in the relatively near
future. I'm less sure about the MinGW version of it though.

GC> But probably we should first build everything routinely with clang.

 I've already started to set up continuous integration, using both gcc and
clang, on Travis CI, which is very easy to use from GitHub. I switched to
more urgent tasks without finishing it (story of my life in a nutshell...),
but I still hope to finish it one day. And then you could receive emails
from Travis whenever any build gets broken (of course, you could also
choose to not receive them but just go to the status page to check the
status manually).

 If you don't object, I'll try to spend some time on finalizing this in
near future. On its own, it's very simple to do, but I'd like to also
create Debian packages for wxPdfDocument to avoid building it every time on
Travis and keep build times shorter (and avoid putting unnecessary load on
free service), and this risks taking slightly more time.

 Regards,
VZ