Darcs checks to see if it can lock WAY to late in the process when pushing.

    $ darcs push [path to remote]
    ...
    [ spend next few minutes choosing patches which I want to push ]
    ...
    Shall I push this patch? (121/121)  [ynW...], or ? for more options: y

    Waiting for lock [path to remote]/_darcs/lock
    Couldn't get lock [path to remote]/_darcs/lock
    darcs: fd:8: hPutBuf: resource vanished (Broken pipe)


Having the user spend time choosing patches carefully one by one and 
then instantly giving up as soon as it can't lock at the end of it 
(rather than saying "hey, I can't lock, try again?" or something) is 
rather poor design.

Attachments

• unnamed

see patch #1333

I added some comments to the proposed solution (patch1333).

The following patch sent by Ben Franksen <benjamin.franksen@helmholtz-berlin.de> updated issue issue2327 with
status=resolved;resolvedin=2.12.0 HEAD

* resolve issue2327: ask user to keep trying if locking fails 
Ignore-this: a2e5649359fc677d2a193c7aa5a65d5c