When I darcs push to a box with darcs 2.2.0 compiled with amd64 (which is, in this 
case, alioth.debian.org), darcs hangs both with 2.2.0 and with darcs-beta 2.2.98.4 
(beta 4).  With 2.2.0 I can kill the local process with C-c, altough a process is 
left in the other box.  With darcs-beta C-c doesn't kill the local process.

Hi Marco,

Do you get the same inability to C-c Darcs 2.3?

What happens if it's a local push? (Darcs push works the same way on
local/remote; it performs darcs apply).

Out of curiosity, do you have any ideas on what could be causing the hang in the
first place?

Darcs developers: it sounds like for testing this, we would need some way of
reliably making Darcs push take a long time, the simpler the better.

Hi Eric.

Em Seg, 2009-08-10 às 16:01 +0000, Eric Kow escreveu:
> Do you get the same inability to C-c Darcs 2.3?

Yes.

> What happens if it's a local push? (Darcs push works the same way on
> local/remote; it performs darcs apply).

I don't have time to test it, since the local push is very fast.

> Out of curiosity, do you have any ideas on what could be causing the hang in
> the first place?

I talked to mornfall on #darcs, and he said he thinks it's the same hang on the
pager, when darcs is build with GHC older than 6.10.2 in 64 bits platforms.
This is the case on alioth, not in my box.

> Darcs developers: it sounds like for testing this, we would need some way of
> reliably making Darcs push take a long time, the simpler the better.

I don't know, but maybe this unability to C-c can be related to the 64 bits
hang.

Greetings.

Thanks for the details.  Do you have the ability to get the folks managing
alioth to install a fresh build of Darcs using a more recent GHC?  Perhaps the
course of action is to try that and see if it goes away?

Hi Eric.

Em Qua, 2009-08-12 às 12:56 +0000, Eric Kow escreveu:
> Do you have the ability to get the folks managing alioth to install a fresh
> build of Darcs using a more recent GHC?

I don't think they'd do it just for testing, but I can reproduce this bug in my
box with a lenny chroot.

> perhaps the course of action is to try that and see if it goes away?

I'll see if this bug is gone with a newer GHC and tell you.

Greetings.
-- 
marcot
http://marcot.iaaeee.org/