Working on implicitly discovered, explicitly built Clang modules at Apple.
- User Since
- Oct 4 2020, 4:28 AM (28 w, 2 d)
This LGTM from command-line perspective, but I'll let others judge the rest of the patch.
Rebase, delete RefBase copy assignment
Thanks for the review!
Fri, Apr 16
Thu, Apr 15
I left a couple of comments inline on the command-line parsing aspect of the patch. For more info, check https://clang.llvm.org/docs/InternalsManual.html#adding-new-command-line-option.
Wed, Apr 14
Sorry, I didn't realize that by "next step" you were referring to a future patch. I'd like to see my two comments addressed in this (and a prep) patch, instead of in the future.
Tue, Apr 13
Remove copy-assignment changes
Mon, Apr 12
Fri, Apr 9
The CodeGen tests you added are failing pre-merge checks. This is most likely because we recently (D97462) started verifying that all all CC1 command line options can be serialized from a CompilerInvocation instance. To reproduce this locally, you'd need to build with assertions or manually set -DCLANG_ROUND_TRIP_CC1_ARGS=ON when building with CMake.
Thu, Apr 8
Thanks for working on this, Paul!
Tue, Apr 6
Back to green after the revert: https://lab.llvm.org/buildbot/#/builders/121/builds/6307
Mon, Apr 5
Fri, Apr 2
Wed, Mar 31
Tue, Mar 30
Sat, Mar 27
Fri, Mar 26
Sorry for taking so long. I created a patch that should fix the missing SourceManager: D99414
Thu, Mar 25
Wed, Mar 24
Benchmark compilation and preprocessing as well
Thanks for the update. Changes touching the command line LGTM, but I'll let others confirm the changes to CodeGen are fine too.
Adding the negative option to the driver makes sense to me. I think we could also simplify the option definitions, see my inline comment.