[Xcb] [PATCH v2] expose 64-bit sequence numbers for XLib
Olivier Fourdan
ofourdan at redhat.com
Mon May 18 00:22:34 PDT 2015
Hi Chris,
> Thank you for the ABI/API test with libxcb and constant xcb/proto.
No problem!
> However we'd need to do another test because
> the primary risk for ABI/API issues are the changes in xcb-proto.
Sure, I was focusing primarily on the 64-bit changes (selfishly, I'd admit easily ^_~).
> (As Uli has mentioned, we should not separate the releases for xcb/proto and
> xcb/libxcb.
> And I agree with him.)
>
> Therefore I suggest to repeat the ABI/API test with comparing the following:
> * last official release of xcb/proto and xcb/libxcb
> * compared with git - heads of xcb/proto and xcb/libxcb
I've done that but it gets slightly more ugly^W verbose... I'll post the results ASAP (just need to redo my setup).
> We can ignore the ABI/API issues for the xinput-extension because
> it was not officially supported in the last release. It was not built by
> default.
>
> Some things to do, too:
> * We should enable the xinput extension in the new release.
>
> * Also, we should include Ulis send_fds patch.
>
> * Check for any other pending patches that fix critical stuff.
>
> * Do some testing with real applications and/or beta-test period.
I also started looking into the Xprint issues that Uli reported (and I can easily reproduce as well, so Uli, nope, you're not alone!) but, err, it doesn't seem to be the related to the proto XML changes. I'll continue my investigations there, as I'd like to understand, but if someone else has an readily available fix, don't let me slow down the resolution.
Cheers,
Olivier
More information about the Xcb
mailing list