- User Since
- Apr 28 2020, 4:57 AM (50 w, 1 d)
Thank you for submitting this @MehdiChinoune ! I'm a bit confused - what's specific about MSVC here and in the failure that you are seeing?
Add one missing test
Rebase on top of main
Mon, Apr 12
Apologies for sending such a chunky patch. Please let me know if you see a good way to split this.
Thank you for addressing my comments @ashermancinelli, this looks really good! One general question, why make everything static? Is there any technical benefit? I feel that it reduces readability a bit. If it's not needed, I suggest removing it.
Fri, Apr 9
Continue debugging debug-measure-parse-tree.f90
Try to diagnose debug-measure-parse-tree.f90
Thu, Apr 8
I've just send an RFC for this change to cfe-dev: https://lists.llvm.org/pipermail/cfe-dev/2021-April/067981.html.
Refine comments, remove code for aliases (which didn't work anyway)
Wed, Apr 7
Add DocBrief descriptions in Options.td
As there have been no new comments and this patch has already been accepted, I decided to merge it without waiting for more reviews. Please leave post-commit comments if I missed anything and I will address them. Thank you all for reviewing!
Hi @ashermancinelli , many thanks for this patch. All tests from the original file are covered and there are also some new tests!
Tue, Apr 6
Remove unused variable
Update f18.cpp and driver-help.f90 to improve testing
Mon, Apr 5
Add more reviewers and remove unrelated changes (apologies for the noise!)
Refine the behaviour when using diag options and add tests
Hi @protze.joachim , thank you for testing this so thoroughly!
LGTM, thank you for addressing my comments!
Thu, Apr 1
Wed, Mar 31
Tue, Mar 30
Thank you for working on this! I would be tempted to make this patch depend on https://reviews.llvm.org/D98657. It's up to you.
I really like how you split your tests into two files:
- werror_scan.f captures warning generated by the prescanner
- werror.f captures warnings from the semantic analysis
In every case you added multiple RUN lines to make sure that the behavior is consistent across multiple actions. I think that that's very useful. Ideally, we'd have one central switch for turning warnings into errors and this would be unnecessary. But we're not there yet. In the meantime, could you add a comment explaining why multiple RUN lines are used?
With this change the corresponding configurations will depend on the worker's internal set-up to use the desired compiler. IMO that's totally fine, provided that:
- linaro-aarch64-flang-latest-clang picks at least clang-11, and
- linaro-aarch64-flang-latest-gcc picks at least gcc-10.
Mon, Mar 29
Make sure that -cpp\-nocpp controls command line macro definitions too
With this patch we can share more tests between the drivers and their behavior with respect to module paths becomes more consistent. Great to see this, thank you @arnamoy10 !
Address PR comments + updated comments + rebase
Interestingly, GCC does not _error_ when using gfortran flags:
$ gcc -ffree-form file.c cc1: warning: command line option ‘-ffree-form’ is valid for Fortran but not for C
I've implemented similar behavior for Clang in https://reviews.llvm.org/D99353. Please, could you take a look and see whether that solves your problem?
Sun, Mar 28
LGTM, thank you for working on this! I've left a couple of comments, but these are [nit]s (feel free to ignore!).
Thu, Mar 25
I've not had a chance to add tests yet. I will add them shortly.
Wed, Mar 24
Rebase on top of D99292
I see that -Werror changes the behavior of the driver in 5 different places. I would hope to see 5 new tests to verify each case.
Thank you for updating this @arnamoy10!
Thank you for updating this!
Removing the release tags makes a lot of sense - the semantics of it mean "test the release branch only" (added here).