[PATCH] D74299: [clang-tidy] extend tests of run-clang-tidy

Wed Oct 21 06:59:14 PDT 2020

aaron.ballman added a comment.

In D74299#2339994 <https://reviews.llvm.org/D74299#2339994>, @AlexanderLanin wrote:

> I cannot reproduce the problem and there is just not enough info to go on.

Ugh, I hate when that happens. :-(

> Is there any way to get anything in addition? e.g. run the same test with `-vv` or add some additional print commands?

We could speculatively commit with additional commands if that'd help you with debugging. We'd be purposefully breaking the build for a few minutes which is never ideal, but if we coordinate properly, it shouldn't be too disruptive.

> Maybe it's some access rights problem, or different version of some non obvious dependency or something like that.
> Maybe it's as trivial as having to use double quotes for `-checks='-*,bugprone-sizeof-container,modernize-use-auto'`.
> It seems run-clang-tidy exits with code 2, but there is no way it would do that. So I'm totally lost.
> http://lab.llvm.org:8011/#/builders/109/builds/690/steps/6/logs/FAIL__Clang_Tools__run-clang-tidy_export-diagnosti

Could it be that some other part of the test is what's generating that result code (like, could it be the result from python of running lit)?

================
Comment at: clang-tools-extra/test/clang-tidy/infrastructure/run-clang-tidy_export-diagnostics.cpp:9
+// RUN: mkdir %t
+// RUN: echo '[{"directory":"%/S","command":"clang++ -c %s","file":"%s"}]' \
+// RUN:      > %t/compile_commands.json
----------------
The `%/S` here looks a bit suspicious to me, was that supposed to be `%S` with the `/` somewhere else?

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74299/new/

https://reviews.llvm.org/D74299