Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

Did not see any failure related to https://reviews.llvm.org/D45936 yet after submission. Best regards Yan Zhang > On Apr 22, 2018, at 18:13, Chandler Carruth wrote: > > I won't be back at a computer for a while and I really don't know anything > about objective-c... But

Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

Sure. Will do. The change I am trying is pretty trivial and only limited to the test itself. On Sun, Apr 22, 2018 at 6:13 PM Chandler Carruth wrote: > I won't be back at a computer for a while and I really don't know anything > about objective-c... But if you don't feel

Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

I won't be back at a computer for a while and I really don't know anything about objective-c... But if you don't feel confident submitting fixes with post commit review, you should probably just revert If the fix is trivial and/or you can find ways to test it a similar suggested in my earlier

Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

Need a accept for that revision. Can you accept it? On Sun, Apr 22, 2018 at 6:02 PM Chandler Carruth wrote: > See my other email -- you can compile code targeting other platforms > regardless of the platform you develop on. Not exactly as good as > reproducing it with a

Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

See my other email -- you can compile code targeting other platforms regardless of the platform you develop on. Not exactly as good as reproducing it with a bot, but about the best you have. On Sun, Apr 22, 2018 at 6:01 PM Chandler Carruth wrote: > I don't know anything

Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

I don't know anything about objective-c, or anything about OSX. However, I guarantee these bots aren't using 32-bit OSX. ;] Look at the bot names. They're running Linux in various flavors: ppc64be, ppc64le, armv8, etc. My suspicion is that this is a linux-specific issue. But you can reproduce

Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

btw due to the lack of way to testing it on bot environments, I am not sure if specify fobjc-abo-version=2 could solve the problem (could it break builds on 32-bit machines or just skip them). So I played safe to use the old way of declaring ivars in the revision. On Sun, Apr 22, 2018 at 5:55 PM

Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

I am running tests locally with "ninja check-clang-tools" and I am sure it is running this test because I could get error message when I debug it. The problem (according to the error message) is all caused by different architecture. It seems a lot of ObjC features are not supported in old 32-bit

Re: [clang-tools-extra] r330559 - update test to use ivar in implementation instead of class extension

The commit log here no longer reflects the commit. This is not just updating the test, this is a complete re-application of the original patch in r330492. =[ Also, the bots are still complaining: http://lab.llvm.org:8011/builders/clang-ppc64be-linux/builds/17830