Skip to content

Add CI using GitHub Actions#3

Open
bonachea wants to merge 2 commits intoBerkeleyLab:developfrom
bonachea:github-ci
Open

Add CI using GitHub Actions#3
bonachea wants to merge 2 commits intoBerkeleyLab:developfrom
bonachea:github-ci

Conversation

@bonachea
Copy link
Copy Markdown
Member

Deploy some basic CI in Actions to serve as a (relatively lightweight) smoke test

@bonachea bonachea requested a review from PHHargrove April 24, 2026 04:09
@bonachea bonachea marked this pull request as draft April 24, 2026 04:21
@bonachea bonachea marked this pull request as ready for review April 24, 2026 16:52
Copy link
Copy Markdown
Collaborator

@PHHargrove PHHargrove left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM.

Curious if --disable-ccs-rpc for Xcode 26 is present for coverage of that case, or if it is necessary on that target as the current comment implies. PR is APPROVED regardless, though an explanatory comment would be appreciated.

@bonachea
Copy link
Copy Markdown
Member Author

bonachea commented Apr 25, 2026

@PHHargrove asked:

Curious if --disable-ccs-rpc for Xcode 26 is present for coverage of that case, or if it is necessary on that target as the current comment implies. PR is APPROVED regardless, though an explanatory comment would be appreciated.

  # XCode 26 / gcc bug workaround: 
  # <mach/mach_time.h> and <mach/mach.h> generate xnu_static_assert_struct_size errors from <mach/message.h>
  if [[ ${{ matrix.os }} =~ macos-26 ]] ; then
    echo "CONFIGURE_ARGS=$CONFIGURE_ARGS --enable-force-posix-realtime --disable-ccs-rpc" >> "$GITHUB_ENV"
  fi

Both configure options are workarounds for the same underlying problem, namely the brew installs of GNU G++ are incapable of compiling the <mach/message.h> header installed on GitHub's macos-26-* runners. Timers reach the failure on the problematic header via include of <mach/mach_time.h> and CCS reaches it via include of <mach/mach.h>.

I feel like the existing comment already conveys this information, but if you disagree then please use the suggest feature to specify your concrete suggested improvement.

The runners were upgraded yesterday with a new version of XCode 26, but I've just re-confirmed here that the problem still exists when we don't disable CCS (and AMUDP fails in a similar way when Mach timers are enabled).

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.

2 participants