propagate git-annex -c on to transferrer child process
authorJoey Hess <joeyh@joeyh.name>
Tue, 15 Dec 2020 15:36:25 +0000 (11:36 -0400)
committerJoey Hess <joeyh@joeyh.name>
Tue, 15 Dec 2020 15:36:25 +0000 (11:36 -0400)
commit74c1e0660b29a293d297d066f147539d548d4dfa
tree16ae0c46ff77bdd9829b325d9636bd10fa17b05a
parent00526a6739db7a8ed95e4f8666eb37045806fb74
propagate git-annex -c on to transferrer child process

git -c was already propagated via environment, but need this for
consistency.

Also, notice it does not use gitAnnexChildProcess to run the
transferrer. So nothing is done about avoid it taking the
pid lock. It's possible that the caller is already doing something that
took the pid lock, and if so, the transferrer will certianly fail,
since it needs to take the pid lock too. This may prevent combining
annex.stalldetection with annex.pidlock, but I have not verified it's
really a problem. If it was, it seems git-annex would have to take
the pid lock when starting a transferrer, and hold it until shutdown,
or would need to take pid lock when starting to use a transferrer,
and hold it until done with a transfer and then drop it. The latter
would require starting the transferrer with pid locking disabled for the
child process, so assumes that the transferrer does not do anyting that
needs locking when not running a transfer.
Annex/Init.hs
Annex/Path.hs
Annex/PidLock.hs
Annex/TransferrerPool.hs
Assistant/Threads/Transferrer.hs
Assistant/TransferSlots.hs
Upgrade.hs
doc/todo/git-annex_transferrer_does_not_propigate_-c.mdwn