llvm: use proper syntax for --config option #61778
Conversation
Use proper --config= syntax instead of --config as the latter can cause issues in some situations. Signed-off-by: Benjamin Cabé <[email protected]>
|
For some unknown reason this just broke the SOF fuzzer build:
Any idea? |
|
Same failure reproduced with more recent clang 15.0.7 @kartben which clang version did you test this with and how? Also, what specific problem(s) did this commit help with? |
clang 16.0.0.0
When using the newly introduced CodeChecker stuff with CodeChecker uses the compile_commands.json file generated by CMake (I think?), which contains entries such as: and for some reason it gets the --config argument wrong? Just adding double quotes around Also, for what it's worth: |
Thanks, that helps.
I found a clang 16.0.6 version somewhere and I tested it and for me it works BOTH with and without the equal sign
I tried to reproduce your issue but I could not. I don't understand why clang 16.0.0.0 seems to drop the config file argument in your case. This feels like a bug in the parser change that they just made, can you try to upgrade your clang? |
Some clang gitarcheology...
I didn't find any obvious bug fix after that. On the other hand, I found some significant changes after 16.0.0, for instance: So I think it's definitely worth trying something newer than 16.0.0.0 |
Also, can you try to re-run your clang command directly on the command line? The easiest to find the real, full command is to temporarily insert garbage in one C file; then CMake prints it and you can copy/paste it and run it directly. This is how I performed all my testing of clang 15 and clang 16. |
This reverts commit 577d47f. The --config= syntax with the equal sign does not work with clang 15 which is less than one year old (https://releases.llvm.org/). It does not work with clang 14 either. ``` clang --verbose clang version 15.0.7 clang --help | grep config -cl-std=<value> OpenCL language standard to compile for. --config <value> Specifies configuration file --cuda-feature=<value> Manually specify the CUDA feature to use clang: error: unsupported option '--config=/home/runner/work/... ``` The reverted commit mentioned "issues in some situations" without providing any example or other information. This revert fixes the SOF fuzzer build, see more details in #61778. Signed-off-by: Marc Herbert <[email protected]>
So the command as found in the compile_commands.json file definitely works, with both syntaxes, and I can make it fail by pointing to a config file that doesn't exist. I am calling for a bug in CodeChecker as it seems to actually be dropping the config file altogether when calling clang-tidy The following from the compile_commands: becomes: |
This reverts commit 577d47f. The --config= syntax with the equal sign does not work with clang 15 which is less than one year old (https://releases.llvm.org/). It does not work with clang 14 either. ``` clang --verbose clang version 15.0.7 clang --help | grep config -cl-std=<value> OpenCL language standard to compile for. --config <value> Specifies configuration file --cuda-feature=<value> Manually specify the CUDA feature to use clang: error: unsupported option '--config=/home/runner/work/... ``` The reverted commit mentioned "issues in some situations" without providing any example or other information. This revert fixes the SOF fuzzer build, see more details in zephyrproject-rtos#61778. Signed-off-by: Marc Herbert <[email protected]>
Use proper
--config=syntax instead of--configas the latter can cause issues in some situations.