On Fri, 3 Apr 2009, Peter W. Draper wrote:
> There must be some further case I'm not trapping when checking the state
> of submodules, so I'll try and have another go at getting the behaviour
> you see. If that fails I'm tempted to remove all the cleverness and
> leave submodule developers exposed to this issue.
Think I've tracked down the mis-behaviour.
When a submodule shows as having modifications I was testing if the
requested HEAD was already in the submodule and using that as an
indicating someone was working on the submodule already. Turns out that
will fail if the submodule has been pulled before the HEAD update, but
after the change has been committed (so in fact deleting and getting a
fresh checkout made this problem worse).
I now check if the actual HEAD of the submodule is ahead of the requested
HEAD instead. So that tests if any commits have not been pushed.
Anyway I expect you'll let me know of any further problems.
Peter.
|