[darcs-users] Re: current status of darcs

Ketil Malde Ketil.Malde at bccs.uib.no
Thu Mar 8 10:29:49 UTC 2007


Simon Marlow wrote:
> Sometimes you don't get a lockup, you get a crash [...] There's no 
> good workaround. 
I think it is clear that darcs has some problems.  I really hope darcs 
won't crumble away because of this and be replaced by something else.  
Isn't it possible to resolve the problem, at least to the extent that 
the pain is bearable?

For instance, is it possible to detect when this situation occurs, so 
that the user gets the option to interrupt and handle things manually?   
More feedback when doing pulls (and similar operations) would be nice 
anyway, I think, and would show the user whether things were progressing 
or not.

(One thing that has landed me into problems occasionally has been 
pulling to a repository with unrecorded changes - a warning with the 
option to bail out/record unrecorded changes would be nice.)

Would it be possible to alleviate the problem by cheating?  Say, 
introducing 'hard' barriers, beyond which no patches would be commuted?  
Or, as somebody suggested, a non-commutable patch type specifying the 
exact new contents of a file? (Such solutions would probably mean 
rejecting patches that do not depend on the barrier, but is still 
preferable to a hanging repository, IMO.)

-k



More information about the darcs-users mailing list