Re: r259862 - [CMake] Improve the clang order-file generation workflow

Actually r259871 includes the code to cleanup the “bad” order file… -Chris > On Feb 4, 2016, at 7:02 PM, Chris Bieneman wrote: > > I pushed a speculative fix for this in r259870. I’m unable to reproduce > locally. There are a few bots that hit the same issue, so I’m watching

Re: r259862 - [CMake] Improve the clang order-file generation workflow

This is causing my links to fail with: FAILED: : && /usr/local/google/home/richardsmith/bin/clang++ -fcolor-diagnostics -std=c++11 -stdlib=libc++ -fPIC -fvisibility-inlines-hidden -Wall -W -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wmissing-field-initializers -pedantic -Wno-long-long

Re: r259862 - [CMake] Improve the clang order-file generation workflow

What linker are you using and on what OS? I’m looking at this now, and will revert if I can’t find a solution. -Chris > On Feb 4, 2016, at 6:38 PM, Richard Smith wrote: > > This is causing my links to fail with: > > FAILED: : &&

Re: r259862 - [CMake] Improve the clang order-file generation workflow

I pushed a speculative fix for this in r259870. I’m unable to reproduce locally. There are a few bots that hit the same issue, so I’m watching them. -Chris > On Feb 4, 2016, at 6:44 PM, Chris Bieneman via cfe-commits > wrote: > > What linker are you using and on

Re: r259862 - [CMake] Improve the clang order-file generation workflow

I'm using gold and it seems like it's been fixed for me. On Thu, Feb 4, 2016 at 10:33 PM, Richard Smith via cfe-commits < cfe-commits@lists.llvm.org> wrote: > This build might be using gold, I don't recall how that machine is > configured. I won't be able to test the fix until tomorrow,

Re: r259862 - [CMake] Improve the clang order-file generation workflow

I reverted the patch and made a pile of other changes to fix up the CMake caches on bots. They should all recover now. -Chris > On Feb 4, 2016, at 8:33 PM, Richard Smith wrote: > > This build might be using gold, I don't recall how that machine is > configured. I won't

Re: r259862 - [CMake] Improve the clang order-file generation workflow

This build might be using gold, I don't recall how that machine is configured. I won't be able to test the fix until tomorrow, hopefully the bits will give you enough. On 4 Feb 2016 7:07 p.m., "Chris Bieneman" wrote: > Actually r259871 includes the code to cleanup the “bad”