The answer to your question is whether a consumer who built against apr<1.7.0 is going to blow up, whether they borrowed "private" API's or not. If they were exported, it's effectively public.
Or, whether a consumer built against 1.7.1 would blow up against 1.7.0 - if that's true, we need to revert. Those are the last of my obvious observations, I'll need to diff the resulting include/... tree between 1.7.0 and 1.7.x branch to know for sure, but thanks everyone for reviewing these questions. Cheers, Bill On Tue, Feb 15, 2022 at 7:24 AM Ivan Zhakov <i...@visualsvn.com> wrote: > > On Wed, 9 Feb 2022 at 13:47, Ivan Zhakov <i...@visualsvn.com> wrote: >> >> On Tue, 8 Feb 2022 at 21:58, Evgeny Kotkov <evgeny.kot...@visualsvn.com> >> wrote: >>> >>> Ivan Zhakov <i...@visualsvn.com> writes: >>> >>> > This part is now in the following branch: >>> > https://svn.ostyserver.net/svn/asf/apr/apr/branches/win32-pollset-wakeup-no-file-socket-emulation >>> > >>> > What do you think? >>> > >>> > It would be great if someone could take a look on the implementation from >>> > the *nix perspective. >>> > After that, I propose to merge the branch into trunk. >>> >>> In case this helps, I have tested the branch on Ubuntu x64 and it seems >>> to compile and pass the tests. >>> >>> >> >> Thanks! Merged branch to trunk on r1897895. >> >> Backporting changes to 1.8.x in my TODO list. >> > Nominated this change to 1.8.x branch. Please review. > > -- > Ivan Zhakov
