GIR isn't compatible with the flags passed for address sanitation and documentation depends on GIR.
Created attachment 456822 [details] Patch
Comment on attachment 456822 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=456822&action=review > Source/cmake/OptionsGTK.cmake:286 > +# FIXME: Does the new documentation toolchain work on Mac now? Do we have any evidence that it does not work in MacOS? TBH, I think the if(APPLE) conditional is unnecessary. It's not like MacOS is particularly well supported, and people packaging for such an environment (e.g. MacPorts) sure can decide themselves if they want to disable documentation or not. FWIW, there is nothing particularly special about gi-docgen: it's a pure Python program, so as long as g-ir-scanner can create a .gir file, it will work fine. Also, this is unrelated to sanitizers, so one more reason to leave this part out od this patch.
(In reply to Adrian Perez from comment #2) ork fine. > > Also, this is unrelated to sanitizers, so one more reason to leave this > part out od this patch. I'm just maintaining this existing behavior. I think someone should double-check that the new documentation tool works on Mac before turning it back on.
(In reply to Martin Robinson from comment #3) > (In reply to Adrian Perez from comment #2) > ork fine. > > > > Also, this is unrelated to sanitizers, so one more reason to leave this > > part out od this patch. > > I'm just maintaining this existing behavior. I think someone should > double-check that the new documentation tool works on Mac before turning it > back on. That said, I don't want to block this patch. I have uploaded a new version that maintains the broken line which tries to disable documentation when building on Mac.
Created attachment 456898 [details] Patch
Comment on attachment 456898 [details] Patch Thanks for the review!
Committed r292528 (249366@main): <https://commits.webkit.org/249366@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 456898 [details].
<rdar://problem/91406165>