Conversation
83429c1 to
d685f05
Compare
Signed-off-by: Rui Chen <rui@chenrui.dev>
d685f05 to
be519f0
Compare
|
Thank you for the MacOS support för building. I would ask if you possible could create a preset and a toolchain file for MacOS in Also, you dont have to refer to previous commit in code. |
|
thanks for the feedback, I am updating the pr now. |
Signed-off-by: Rui Chen <rui@chenrui.dev>
|
Followed up on this in Changes made:
I also set |
rampageservices
left a comment
There was a problem hiding this comment.
A few observations from reviewing:
_GNU_SOURCE in cmake/toolchains/macos-clang.cmake — consistent with the Linux toolchain which also defines it, so this makes sense for parity.
continue-on-error: true on the test step — is this intentional as a temporary measure while macOS tests stabilize, or should it stay permanently? If temporary, it might be worth adding a comment noting why, so it doesn't get forgotten. If permanent, it means macOS test failures will never block CI, which reduces the value of running them.
CMAKE_COMPILE_WARNING_AS_ERROR: OFF — the Linux preset doesn't set this at all (inherits whatever the project default is). Is macOS hitting warnings that Linux doesn't? If so, it might be worth documenting which warnings are macOS-specific, since silencing them project-wide could mask real issues on that platform.
bzip2 pkg-config file generation — this is the standard approach for Homebrew keg-only packages that don't ship .pc files. Looks correct.
Job rename from ci to ci-ubuntu-22.04 — worth double-checking that no branch protection rules reference the old ci status check name, since GitHub treats these as distinct checks.
actions/checkout@v6 — consistent with the existing Ubuntu job. Neither is SHA-pinned, which is fine for a project of this size, but noting it for completeness.
Overall looks solid as a smoke-test addition. The main question is whether continue-on-error should remain long-term.
|
Thank you for the I just checked the CI build and it reported the following error https://github.com/gerbv/gerbv/actions/runs/22560526941/workflow Seems there can not be number in ci work names. And also the comnment above by @rampageservices on at least commenting if it is temporary or permanently off. It is great to test it on Clang as well and (later) making sure it is warning free there as well. But it is a journey... |
|
yeah, that should easy fix (I ran into that issue the other day, that the job id does not support dot in |
Signed-off-by: Rui Chen <rui@chenrui.dev>
|
@spe-ciellt should be good now |
|
Thanks for the updates. OK, I see the warnings from Clang now. Interesting, Clang and GCC seems to disagree how much I will merge this now. |
|
Thanks @spe-ciellt. |
Add a native macOS CI smoke build on
macos-15for CMake/Homebrew validation.Follow-up to #303.