[llvm-branch-commits] [mlir] [MLIR][Standalone] test Standalone against install distributions (PR #157944)

[Maksim Levental via llvm-branch-commits]

makslevental wrote:

> The fact that Subprocess 1 CMake and Parent CMake have identical build directories seems particularly problematic.

My assumption was (and I realize that it's flawed now) that process 1 isn't actually building anything, just "installing" artifacts. But of course we've all seen things get built when doing `ninja install` even after doing a full build.

> Sounds like we're aligned that this Standalone/LIT location isn't the right place, @makslevental ?

I mean I don't care how the thing is tested right - I was just going for what I thought was the least controversial approach. IF there's an even less controversial approach I'm happy to do that instead!

BTW I guess this is what @boomanaiden154 was talking about

https://github.com/llvm/llvm-project/blob/e236a52a88956968f318fb908c584e5cb80b5b03/libcxx/test/CMakeLists.txt#L40-L58

which I can try out if we _do_ want it to be a lit test (but again happy not to have to do that).



https://github.com/llvm/llvm-project/pull/157944