Thanks for checking this out, TWB.  When I tried the same experiment  
I got a 14-minute runtime.  Then I ran "darcs optimize --checkpoint"  
on the server, and tried again and it was still 14 minutes.  Then I  
added "--partial" to the "darcs get" command-line and my local  
workstation locked up.  :-(  I suspect that darcs tried to use all  
the physical RAM in the machine (4 GiB) and then some and thrashed  
until Linux killed it.

I tried it again and watched "top" and sure enough, that's what happens.

So the bad news is that darcs-1-format repositories are unusable for  
allmydata.org, even using darcs-2.1.2 executables.

The good news is that I then tried "darcs get --lazy http:// 
allmydata.org/source/tahoe/trunk-hashedformat tahoe" and that worked  
much better.  It took 1m35s, and didn't consume incredible amounts of  
RAM.

The upshot is that I think the time has come to force my users to  
upgrade to darcs-2 and hashed-format repos.  :-/

I'm not looking forward to that, in part because someone might say  
"Hey, if we're switching revision control tools, why not switch to  
git?".  I'll let you know how it goes.

Regards,

Zooko
---
Tahoe, the Least-Authority Filesystem -- http://allmydata.org
store your data: $10/month -- http://allmydata.com/?tracking=zsig
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users

Reply via email to