- User Since
- Jun 10 2016, 1:55 AM (110 w, 1 d)
Fri, Jul 6
I don't really know anything about the AMDGPU target, so I'll resign as reviewer to make that clear, but thanks for notifying me about the fix!
Thu, Jul 5
Hi! We encountered a UBSan runtime error after this was merged. I wrote a bug report about it: https://bugs.llvm.org/show_bug.cgi?id=38071.
Mon, Jul 2
Thanks for the reviews!
Wed, Jun 27
Use pop_back_val() for getting the items from the worklist.
Jun 20 2018
May 31 2018
May 30 2018
Any more comments or concerns, or can I land this?
May 29 2018
May 28 2018
Addressed vsapsai's comments.
Update patch to include the clang-tools-extra test case originally added in D47251.
May 25 2018
Query TheDriver.isSaveTempsEnabled() at uses instead of storing the value in the constructor.
Renamed SaveTempsEnabled field to KeepTempFiles.
I have now updated the patch so that the files are removed when deleting Compilation objects.
May 24 2018
Thanks for the review! I'll submit this tomorrow when I have time for the build bots.
May 23 2018
We have added a lit reproducer for this now in clang-tools-extra: https://reviews.llvm.org/D47251.
May 22 2018
May 18 2018
Thanks for the reviews! I'll submit this shortly.
May 17 2018
Any more comments on this?
May 10 2018
May 9 2018
@bruno, @vsapsai: I added you since you I saw that you recently reviewed, respectively delivered, D30881. That is the only DependencyFile commit since October; although, this feels a bit orthogonal from that, so feel free to remove yourselves as reviewers (and I'm sorry for adding you in that case)!
Apr 25 2018
Ping. It feels a bit nasty that the tools leave behind temporary files, so I think that it would be good to find a fix for that.
Apr 18 2018
Apr 16 2018
Apr 11 2018
Hi! We have encountered a regression where clang-tidy leaves behind temporary files after this change. I wrote a PR for that: https://bugs.llvm.org/show_bug.cgi?id=37091.
Mar 28 2018
Our legacy frontend does not support -MJ, so when using that frontend for code generation, we also invoke clang with -MJ, and at the same use -fsyntax-only to get the improved diagnostics that clang provides. This is idiosyncratic and probably hacky, I know, but it works well enough to for example for getting access to defines and include flags from the compilation database, and being able to run clang-tidy. So (1) does not fit our use case, unfortunately.
With this change, we will emit the .file 0 directive even for -gdwarf-. The directive results in an error when assembling with GAS. Should this be seen as a problem, or is it something that we accept?
Mar 26 2018
Downstream we use -MJ in a bit of an idiosyncratic way, as we're in a transition period where we, for a subset of the code base, only use the clang frontend for diagnostics, and not for the code generation. However, if you don't think that using -fsyntax-only and -MJ makes sense in any upstream application, I'll drop from this change. I'm leaving the assertion as-is.
Mar 22 2018
Mar 16 2018
A small caveat with this patch is that it does not fix the case where the input file as also added as an extra dependency with -fdepfile-entry; however, I reasoned that it shouldn't really be a problem in practice. I thought that it was a good trade-off ignoring that for slightly simpler code.
Mar 2 2018
Feb 1 2018
Removed nonessential attributes.
Jan 26 2018
Jan 25 2018
Jan 24 2018
Oct 18 2017
Unless I'm mistaken, int_experimental_gc_statepoint should also be marked recursive?
Sep 26 2017
Simplified the test case according to dblaikie's comment (thanks!).
Sep 25 2017
Sep 22 2017
Aug 23 2017
Changed so that the debug location is kept in the other cases.
Added CHECK statements (thanks for pointing that out!).