Re: [msysGit] Re: [Tortoisehg-discuss] Bazzar stratgy regarding shellextension

Just to interject; how do you imagine to cache source-code file-statuses
for checkouts that are distributedly-source-controlled, without making a
ridiculous and constantly-high number of network calls to other members of
the project that have their own copies?

In the early days of TSVNCache it was actually quite unreliable because fundamentally
an SVN WC is a shared resource to which access is granted _without locking_.
 I could do a non-SVN operation to a file inside of a WC and my recursive
icon-status in another explorer window might not update for a while, since
there was no inter-proc communication, and this was before the cache had
a file-system watcher (if memory serves).

Distributed source-control seems to be the same problem but now across network
boundaries instead of being merely confined to one box.

It seems like a _hard_ problem to solve! <-- red rag... ;-)

Pete

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_tortoisesvn.tigris.org
For additional commands, e-mail: dev-help_at_tortoisesvn.tigris.org
Received on 2008-04-22 14:49:39 CEST