Introduce Mise #132
Introduce Mise #132
Conversation
nettle
force-pushed
the
multiconfig-test-with-mise
branch
3 times, most recently
from
2cdddf8 to
bffea9e
Compare
Why: We should run all tests for different configurations: - Bazel versions 6, 7 and 8 - CodeChecker versions 6.25, 6.26, 6.27 - probably different Python versions - might be different Clang and Clang-Tidy versions It would be also good to install and use dev tools like pylint, pycodestyle, buildifier What: - Introduce Mise to support multiple configurations - Add example Mise configs for Bazel and CodeChecker - Add script which installs, sets up and runs all configs
nettle
force-pushed
the
multiconfig-test-with-mise
branch
6 times, most recently
from
e952756 to
31b398e
Compare
nettle
force-pushed
the
multiconfig-test-with-mise
branch
from
31b398e to
cab46b2
Compare
There was a problem hiding this comment.
I think we should create an issue where we first debate what tool we want to end up using. Mise is promising, but I'm not sold on it yet. For instance, if I see a failure and kill the process prematurely, the error logs are all lost. But I'm not opposing!
On a more practical note, when running mise, I can't reproduce behaviour known to work without it. I know for a fact that bazel 6 / bazel 7 runs with CodeChecker 6.25 and 6.26, yet the tests still fail when I'm running them through mise. Inspecting the logs reveals this CodeChecker internal error:
[WARNING 2026-01-21 13:06] - 'diagtool' can not be found next to the clang binary
Diagtool is known to be quite tricky to acquire from package managers, so we need to figure out how mise can get that. Interestingly enough, bazel 6 tests pass for me fine, only bazel 7 breaks.
2 participants
Why:
We should run all tests for different configurations:
It would be also good to install and use
dev tools like pylint, pycodestyle, buildifier
What: