@bryanwweber ... thanks for the suggestion!

Interestingly, the hash in the build logs is not the same as the commit shown in the PR here:

INFO: Building Cantera from git commit '4fd09b3'

From https://github.com/Cantera/cantera/runs/4859330648?check_suite_focus=true#step:7:9

Interestingly, the hash in the build logs is not the same as the commit shown in the PR here:

This is truly odd, as I’m just displaying a value that had existed before. No idea why the build creates a different hash (4fd09b3).

PS: Pasting the hash actually resolved the riddle - the commit is a fix-up that was merged into what is now ae316de. I believe this is due to running git commit --amend --no-edit. Funny that git discovers a different hash.

The approach followed in SConstruct to get the git hash (git rev-parse --short HEAD) is the recommended solution on SO. On my machine this works as expected (and is consistent with the output of all other alternatives). So while the build currently displays a different hash, I do not believe that there is an actual issue here.

🎉 FWIW, the merged build displays the correct hash, i.e.

INFO: Building Cantera from git commit 'a0b62a4'

so my interpretation is that the hash mismatch observed above is likely due to some GH internals.