[darcs-users] darcs patch: fix wrong ../path in failing issue1013 test

Trent W. Buck trentbuck at gmail.com
Sun Oct 5 02:47:24 UTC 2008


David Roundy <droundy at darcs.net> writes:

> On Sat, Oct 04, 2008 at 06:50:23PM +0100, Eric Kow wrote:
>> On Sat, Oct 04, 2008 at 19:29:03 +0200, Tommy Pettersson wrote:
>> > I'm mystified by the 'exit 1 # darcs hangs here' line. How did
>> > the test ever get to that line?
>> 
>> Hmm, for what it's worth, I can no longer reproduce this hang with
>> darcs 2.0.3pre1 (with or without your patch).
>> 
>> I would have to try it with darcs 2.0.2+77 to see.  It would be nice
>> if there was an easy way for me to pull the first N patches in darcs
>> just for this one purpose :-D
>>  (I suppose I could do something like automatically saying yes 77
>>   times)
>
> Something like darcs pull -n 30-100 should work, you'd just have to count
> backwards.

"darcs pull --help" doesn't mention any -n switch in 2.1.0pre2.  Is this
a bug in the documentation?

This is what I do.  It's ugly and wrong, but it works.

    echo yyyyyyyyyyyyyyyyyyyyd | darcs pull

In bash or Emacs, you can type the large number of y's with "ESC
<number> y".

Similarly, I use this to push all patches that don't introduce a
conflict (from memory):

    x=yd
    while while echo $x | darcs push; do :; done
    do x=n$x
    done

To improve speed, I can crank up x=yd to x=yyyyd and then once that
always fails, set x back to x=yd to catch the finer grains.

Should darcs return a different exit status if 1) you say "n" (or "q")
to all patches; vs 2) it aborts to avoid merge conflicts?  If it
did/does, the above push loop can be taught to exit when there's nothing
left to push.



More information about the darcs-users mailing list