[darcs-users] Initial checkout is very slow

[Erik Schnetter]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I have just created a darcs repository (converted from CVS; there are 
about 800 changesets in it) and put onto a server.  While getting this 
repository onto my local machine over a DSL line, I find that this is 
very slow.

"darcs get" gets about one patch every three seconds.  Each of these 
patches is less than about 1 KB long.  My local machine shows 
essentially no load, and the network does not seem to be maxed out.

I used the command

	darcs get darcs at cvs.carpetcode.org:/home/darcs/carpet

and I use ssh keys for authorization.

Is it, maybe, the case that darcs opens a new ssh connection for each 
patch?  That would explain the speed.  Things can usually be much 
faster by copying all files through a single ssh connection.

The darcs source code says here essentially " sequence_ ... 
copyFileOrUrl", indicating that each patch is copied sequentially.  The 
curl documentation says that "Curl will attempt to reuse connections 
for all URLs specified on the same command line/config file".  Would it 
be worthwhile to introduce a routine "copyFilesOrUrls"?

What do other people do here?  Copy the repository by hand initially?

- -erik

- -- 
Erik Schnetter <schnetter at aei.mpg.de>   http://www.aei.mpg.de/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from www.keyserver.net.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFBTCIim3uiSwno3f0RAiQFAJ9MZ8wC0MA+X4PTqDlpf2fa8+1wQgCgxvuU
hAfRHK9Jq495Lnq3DwI+cL0=
=tV2l
-----END PGP SIGNATURE-----