That's just on the try bot, on a try run for https://codereview.chromium.org/1828543003/ , right? (my cl)

Huh, no.

clang-cl is supposed to ignore that flag (we currently pass -Qunused-arguments), but it looks like it isn't. Not sure why. I'll add it to the ignored flags upstream.

Since nothing's urgently motivating adding /d2FastFail, probably best to revert that Cl and reland it in a few days after the next clang roll.

…or add clang==0 / !is_clang checks before adding that flag until we rolled in a version of clang that silently ignores that flag.

Fixed upstream here: http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20160328/154086.html

Bruce, wanna revert or add !clang checks until that rolls in?

Revert: https://codereview.chromium.org/1843783002/

clang-cl now knows this flag, and bruce readded it yesterday.

I recently updated my repository to top of tree, and I can't seem to get rid of this error.

From my goma terminal: E0415 ... ... include_processor.cc:634 root include:\d2FastFail not found

oh noes! goma probably doesn't support this in clang-cl mode yet? is your goma client up to date? the bots seem happy.

'GYP_DEFINES': 'gomadir=c:/goma/goma-win64 use_goma=1 component=shared_library',
'GYP_GENERATORS': 'ninja',

Reverting bruce's CL fixed my problem locally. robliao@ said he didn't have this problem.

I didn't think I was using clang...I guess my goma client could be out of date? It automatically starts itself whenever I compile.

python c:\goma\goma_ctl.py update

Oh, if you're not using clang, then yes, your goma client is seriously out of date. It was added there long ago.

goma doesn't automatically update so it's pretty easy to end up way behind. I think the goma team is thinking about ways to address this, but for now you should update as thakis@ suggested.

Thanks. I'm too used to having nice things.