Skip to content

Build the strip.py output in CI - #8

Draft
Stuartwastaken wants to merge 1 commit into
fix/strip-ifndef-unit-testfrom
ci/strip-output-build
Draft

Stuartwastaken wants to merge 1 commit into
fix/strip-ifndef-unit-testfrom
ci/strip-output-build

Conversation

@Stuartwastaken

Copy link
Copy Markdown
Owner

__STRIP__=1 strips with the preprocessor, so nothing in CI compiles what util/strip.py actually writes for upstreaming. This adds a step that runs make net-next into a temp dir and builds the result against the same 6.17.8 tree with -Werror. To make that possible, strip.py --kernel X.Y.Z keeps the code for that Linux version in LINUX_VERSION_CODE blocks instead of always the #else; without the option the output is unchanged.

It needs the strip.py fix in the base branch: without it, this step fails on main today with redefinition of 'homa_tx_pool_mem_nid'. With the PlatformLab#126 strip.py fix (dc2f395) reverted, it fails with homa_timer.c:106: suggest explicit braces to avoid ambiguous 'else' (https://github.com/Stuartwastaken/HomaModule/actions/runs/40747769139).

__STRIP__=1 strips with the preprocessor, so nothing compiled what
util/strip.py writes, and mistakes like the dangling else fixed in
dc2f395 went unnoticed. strip.py --kernel keeps the code for a given
Linux version in LINUX_VERSION_CODE blocks, so the output builds with
the CI kernel.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant