Since WebKit's Python scripts only support 2.7.x, CMake should fail immediately upon discovering 3.x.x -- we shouldn't let the build run for potentially tens of minutes before a syntax error is finally encountered.
Created attachment 306156 [details] Patch
Comment on attachment 306156 [details] Patch This seems like a good idea! r=me.
Comment on attachment 306156 [details] Patch Clearing flags on attachment: 306156 Committed r214886: <http://trac.webkit.org/changeset/214886>
All reviewed patches have been landed. Closing bug.
Why don't we just fix the problems with python 3 and support it? I think that's a better plan long-term.
For sure. I figured this was a less controversial change to start with though, and might get more people to ask the question you just asked. :P I'd be happy to update the scripts as a follow-up task, but I'm worried about how to ensure they *stay* compatible thereafter...?
I could uninstall python2 and install python3 on at least my bot on build.webkit.org once it works, and then if it breaks the breakage would be noticed.
(In reply to Alex Christensen from comment #7) > I could uninstall python2 and install python3 on at least my bot on > build.webkit.org once it works, and then if it breaks the breakage would be > noticed. There are automated tools we can run scripts through https://docs.python.org/3/howto/pyporting.html. There is a bug for Python 3 support at https://bugs.webkit.org/show_bug.cgi?id=156674
Could it be sustainable to support python2 and python3? If not, then we should just require python2
(In reply to Alex Christensen from comment #9) > Could it be sustainable to support python2 and python3? If not, then we > should just require python2 It sounds like its doable. I'm just not sure how much of a change it will be in the existing python code base. We can try running some of the automated tools over the tests and see how much the code changes.