- User Since
- Jan 10 2013, 2:43 PM (322 w, 6 d)
Thanks, landing with UNSTABLE part removed.
Done, thanks for catching that. As discussed on chat, the patch description already had a description on why this is a problem :)
Made obsolete by D57609
Tue, Mar 19
Mon, Mar 18
Ah, it was here because when rL288992 was written we still used UnDecorateSymbolName() for demangling. Now we use LLVM's microsoftDemangle() which is cross-platform.
Fri, Mar 15
Thu, Mar 14
Why is this needed for index-while-building? My mental model for index-while-building is that that streams out build index metadata as part of the regular compile. Why does that require watching directories?
ilya-biryukov, ping? :-)
Wed, Mar 13
Ah, got it, thanks for explaining!
Seems like this wasn't convincing, so now with two separate feature flags. I went with the shorter name, since adding "interface_" isn't just longer but also misleading, since this also works in @implementation ivar blocks now.
Note to self, this was for https://crbug.com/751225
Mon, Mar 11
(A few python style comments; feel free to ignore, and feel free to land no matter what you do with the comments.)
Change lld-link instead.
(Since jfb added a few people who might not have seen it: This was discussed in "Objective-C++11: concerns about adding static_assert support to @interface / @implementation?" on cfe-dev)
This breaks building with clangs that don't have https://reviews.llvm.org/rL267957 , such as Xcode 8; lots of errors like
Fri, Mar 8
Ah ok, then I agree this doesn't need serialization yet :)
I landed the follow-up to the shell=True change in r355694 since it wasn't done yet and I had three spare minutes this morning.
Thu, Mar 7
Does our serialization machinery serialize these correctly automatically? What happens if you compiler a header saying execution_character_set_push, "utf-8") in a header, compile that to a pch, then include that through the pch in a .c file that does pop?
Tue, Mar 5
Thanks for the patch1
Mon, Mar 4
Sat, Mar 2
Thu, Feb 28
Wed, Feb 27
Looks like this broke most build bots, see http://lab.llvm.org:8011/console
Tue, Feb 26
Mon, Feb 25
Fri, Feb 22
If we do this, should we do it in lit? Else we probably need to do it in llvm, clang, lld, clang-tools-extra, etc etc etc.
Thanks, but looks like @hctim did this in r354683 already (thanks hctim too)!
Wed, Feb 20
Hmm, that's right. That's probably a good reason why we can't make P the default behavior.
Tue, Feb 19
But D57425 landed and this one didn't :-) Is the plan to land this and revert the other one then? It sounded to me like our libc++/win bot is currently happy, what does this change do that the other one misses?
This was made obsolete by D57425, right?
Feb 19 2019
Feb 17 2019
Feb 15 2019
Feb 14 2019
Feb 13 2019