> > > On 8 Mar 2018, at 12:42, Christophe Fergeau <cfergeau@xxxxxxxxxx> wrote: > > > > On Thu, Mar 08, 2018 at 05:39:48AM -0500, Frediano Ziglio wrote: > >>>> There are however still some issues: > >>>> - the syntax is using C++20 while we state we use C++11 syntax, this > >>>> is basically using C compatibility extensions. I just tried and for > >>>> instance this code is not accepted on Visual C++ 2015 (not an issue > >>>> at the moment); > >>> > >>> No, but it is annoying. Will make that obvious in the commit log. > >>> > >> > >> I don't think that a comment on the log will make Visual C++ > >> compile that code. Stating C++11 was the reason of this, not use too > >> advance syntax that could have problems. > > > > For what it's worth, I'd be in favour of *not* using things newer than > > what is in c++11 (or maybe c++14), this would give some first guideline > > as to what's ok to use, and what should be avoided. > > I agree. > > However, we presently use gnu++11, so this is C++11 with GNU extensions. > As an extension, designated initializers have been in GCC since at least 4.7. > > So do you think we should avoid GCC extensions and switch to “strict” mode? > > > Christophe No, you don't agree, C++11 is not gnu++11. IMHO your argument is OT. There's a difference in including external code that uses extensions or use extensions in a portable way if strictly needed or not breaking portability. Frediano _______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel