Network performance problems when pulling and cloning from HTTP server

[Pierre-Yves David]

On Tue, Nov 27, 2012 at 12:54:51PM +0100, Angel Ezquerra wrote:
> On Tue, Nov 27, 2012 at 12:24 AM, Matt Mackall <mpm at selenic.com> wrote:
> > In theory, Mercurial could lazily read the incoming bundle and only read
> > past the changelog portion if we actually needed that data, but it turns
> > out we don't actually do this yet. So incoming is currently exactly as
> > expensive as pull.
> 
> I see. Thanks for the explanation.
> 
> TortoiseHg always saves incoming bundle, and then uses it if you do a
> pull after incoming. Thus in that case the cost is not very high.
> However, what do bare mercurial users do?
> 
> I guess one solution would be to create an alias that saved the
> incoming bundle into some "last.bundle" file and then another alias
> which would pull from "last.bundle"... Is there any other good
> solution?

The current bundle format does not contains phase data. This pull by incoming +
pull from bundle lack bookmarks and phases data.


-- 
Pierre-Yves David

http://www.logilab.fr/

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20121127/e0c1663a/attachment.asc>