Jason Bacon <jtocino%gmx.com@localhost> writes:
So, for the branch, have we identified a defective sdk, or a defective
12, such that we should in some mk file, conditioned on
macos
sdk
?
set that, so that building on macs will work?
Do we want to tamper with the mk files, given that this is likely a
transient problem with an easy workaround? This could be fixed by
patching gcc12.
Hard to say.
If it is fixable, it would be great to fix it. Changing to 13 for
(some?) macos seems lower risk to everytbody else.
The extra patches in wip/gcc12 were working for me a couple weeks ago,
but they are not at the moment. I can dig a little deeper if there's a
motivation to fix this before the branch.
In general we like the branch to work :-) Who is motivated to fix what
is of course another question! So if you are up for figuring it out,
great, but we have to be careful not to cause regressions on platforms
that people won't be able to test on, so I am reluctant to land patches
into gcc unless everybody thinks they are clearly right with low risk.