Added support for Texas Instruments C6000 compiler.

[Machapet]

Created new compiler and linker classes:

• C6000CCompiler
• C6000CPPCompiler
• C6000Linker
• C6000DynamicLinker

It was necessary the compiler can detect both extension for static libs (*.a, .lib) and for dynamic libs (.dll, *.so) => added function is_baremetal()

Added new cpu family 'c6000' into known_cpu_families for which the C6000 compiler is used (https://www.ti.com/tool/C6000-CGT)

[tristan957]

Please squash the commits too!

[tristan957]

This looks pretty clean. I marked it on the 1.3.0 milestone, so hopefully someone with merge powers can come through and click the button!

[tristan957]

Would you mind adding a release snippet?

[Machapet]

Would you mind adding a release snippet?

Well... but how to do it? Just write here into this conversation? Or should I do it some special way? I'm sorry, I'm new to these processes on github :-).

[tristan957]

@Machapet look in docs/markdown/snippets and add a new file there.

[mon]

I see you've copied the comment I left on the C2000 compiler:
# Required for backwards compat with projects created before ti-cgt support existed

Ideally, all TI compilers should now be pretty much supported, though I see you probably have to add the compiler .exe name at least. Does it work if you use the TI armclang PR I made in #11421 (which in an ideal world would also make it into 1.3) ?

[mon]

You may also need to combine it with my other pending PR, #11942 - which removes the linker name requirement entirely (compiler name is still needed for detection though, which could be fixed)

[Machapet]

I see you've copied the comment I left on the C2000 compiler: # Required for backwards compat with projects created before ti-cgt support existed

Ideally, all TI compilers should now be pretty much supported, though I see you probably have to add the compiler .exe name at least. Does it work if you use the TI armclang PR I made in #11421 (which in an ideal world would also make it into 1.3) ?

I'm not sure if I understand what I'm expected to do. Currently we are not using TI armclang as it was not suitable for us, but neither me nor my collegues don't remember what was the reason, it is more than 1 1/2 year. We are also using TI-CGT v20.0.6.LTS (armcl, armar...) for some AM1X based devices (TI ARM) and there were no problems to use Meson with this compiler. Have I answered your question?

[mon]

I'm not sure if I understand what I'm expected to do

Sorry for the confusion. If the c6000 compiler must be different from the "generic" TI-CGT implementation (i.e. your modifications to stlibext and shlibext are required for it to work), then you should remove the comment # Required for backwards compat, since this is incorrect.

If the c6000 compiler can be treated the same as other ti-cgt compilers, then I think it would be good to remove the C6000 classes (CCompiler/CPPCompiler/Linker/DynamicLinker) and simply use the TICompiler classes.

The misidentification of TI armclang that you moved the code to line 442 for should be solved by my armclang support PR, though it shouldn't really change your use of ti-cgt.

[Machapet]

I'm not sure if I understand what I'm expected to do

Sorry for the confusion. If the c6000 compiler must be different from the "generic" TI-CGT implementation (i.e. your modifications to stlibext and shlibext are required for it to work), then you should remove the comment # Required for backwards compat, since this is incorrect.

If the c6000 compiler can be treated the same as other ti-cgt compilers, then I think it would be good to remove the C6000 classes (CCompiler/CPPCompiler/Linker/DynamicLinker) and simply use the TICompiler classes.

The misidentification of TI armclang that you moved the code to line 442 for should be solved by my armclang support PR, though it shouldn't really change your use of ti-cgt.

I'm afraid I cannot remove my "C6000" classes. I tried to apply your changes into my code without my classes and did not work:
Unable to detect linker for compiler cl6x -Wl,--version. So I remove the comment # Required for backwards compat

[Machapet]

I pushed minor changes (only remove some #comments) and some checks did not pass (https://github.com/mesonbuild/meson/actions/runs/6529125143/job/17726319329?pr=12246, https://github.com/mesonbuild/meson/actions/runs/6529125144/job/17726319000?pr=12246). Where is the problem?

[tristan957]

@Machapet the CI is flaky. Your changes probably didn't do anything to cause the failures.

[tristan957]

Pylint job is failing. Also squash your commits.

[Machapet]

Pylint job is failing. Also squash your commits.

Af for the squash... The commit b1dd1b0 is a result of automatic merge with master (there were some merge conflicts I have already resolved). So should I really squash the 9869131 and b1dd1b0. The last commit contains all the changed from master...

[eli-schwartz]

Are you going to remove the merge commit?

[dcbaker]

Can you rebase this? I think this has been reviewed and can be merged

[Machapet]

Can you rebase this? I think this has been reviewed and can be merged

I'm sorry but what I'm expected to do now? I'm still not familiar with the github workflow. I tried to follow the instructions from here: https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/incorporating-changes-from-a-pull-request/merging-a-pull-request but I can't see any options at my pull request.

[an-dr]

Can you rebase this? I think this has been reviewed and can be merged

I'm sorry but what I'm expected to do now? I'm still not familiar with the github workflow. I tried to follow the instructions from here: https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/incorporating-changes-from-a-pull-request/merging-a-pull-request but I can't see any options at my pull request.

Hey! It's super simple:

1. Sync your fork's main
   https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

2. Rebase to the updated main.