Open jbouzanquet opened 11 years ago
Yes to both, I think. I'd already given some thought to putting the GIT commit number and anything else into the log at the start of the run... that way if we wanted to go back to that code snapshot we could.
That would require typing the commit hash into the file, which would cause a new commit, changing the hash? Hashes are stable, but are the numbers stable even when merging branches?
The commit numbers should be stable, yes... and no, it wouldn't if all we did was run 'git status' from within the script and capture the output and append that to the log. It would tell us exactly where we are with it when it was run.
That works if we assume we're always running on linux. It seems like several things work that way, maybe we should just roll with it.
I'm gonna hammer on this a bit RSN...
Should we label the programs with version numbers, and if so should these be output in log files?