You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Ashish Utagikar <ap...@yahoo.com> on 2008/12/02 23:09:13 UTC
Data caching functionality for subversion
Hi All,
Does anybody know whether subversion has any data caching functionality. By data caching, I mean the data is cached in the cache area so that the the client does not have to contact the server every time and it can get its data from the cache which usually resides on the same machine/file system as the client.
Many tools like Design Sync, SOS etc have this functionality.
Right now it looks like the client has to contact the server everytime which might reside on the remote machine during the update
thanks
Ashish Utagikar
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=978584
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].
RE: Data caching functionality for subversion
Posted by Bob Archer <Bo...@amsi.com>.
I'm a little confused as to what you want. "Update" means get the latest
revision from the repository. I'm not sure how you would update from a
cache?
Or are you talking about using something like svnsync to create a mirror
of a remote repository on a machine on the local lan?
BOb
-----Original Message-----
From: Ashish Utagikar [mailto:aputagikar@yahoo.com]
Sent: Tuesday, December 02, 2008 6:09 PM
To: users@subversion.tigris.org
Subject: Data caching functionality for subversion
Hi All,
Does anybody know whether subversion has any data caching
functionality. By data caching, I mean the data is cached in the cache
area so that the the client does not have to contact the server every
time and it can get its data from the cache which usually resides on the
same machine/file system as the client.
Many tools like Design Sync, SOS etc have this functionality.
Right now it looks like the client has to contact the server everytime
which might reside on the remote machine during the update
thanks
Ashish Utagikar
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageI
d=978584
To unsubscribe from this discussion, e-mail:
[users-unsubscribe@subversion.tigris.org].
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=979189
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].
Re: Data caching functionality for subversion
Posted by Ryan Schmidt <su...@ryandesign.com>.
On Dec 3, 2008, at 18:23, Ashish Utagikar wrote:
> So is there any plan to implement this feature in svn in the future
> or as you said since subversion is designed for low bandwidth
> situations, the developers are not going to do it.?
>
> Hi Bob,
> As Ryan explained when I meant data caching, i mean the
> client has to contact the repository everytime we do svn update,
> svn commit etc.. which might be a hit on the network bandwidth.
> Instead of contacting the repository evrytime, we can get the file
> changes from the cache which lies on the same machine/filesystem
By definition it would not be possible to implement "svn update"
without contacting the repository, since its purpose is to get new
changes from the repository. "svn update" is of course efficient in
that it only gets from the server the difference between what the
client has and what the server has; if they're already identical,
then very little data has to go over the network to determine that
(though it may take some time while the client figures out the state
of the working copy, depending on the size of the working copy).
Though again I encourage you to have a look at svk which is designed
even more for those wanting to work offline. I understand that it
lets you make commits to your own personal copy of the repository,
and then later send those back up to the main repository (assuming of
course there are no conflicts with commits others have already made
to the main repository).
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=979225
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].
Re: Data caching functionality for subversion
Posted by Ashish Utagikar <ap...@yahoo.com>.
Thanks Ryan.
So is there any plan to implement this feature in svn in the future or as you said since subversion is designed for low bandwidth situations, the developers are not going to do it.?
Hi Bob,
As Ryan explained when I meant data caching, i mean the client has to contact the repository everytime we do svn update, svn commit etc.. which might be a hit on the network bandwidth. Instead of contacting the repository evrytime, we can get the file changes from the cache which lies on the same machine/filesystem
I am not talking of client credentials/password caching etc
thanks
Ashish
--- On Wed, 12/3/08, Ryan Schmidt <su...@ryandesign.com> wrote:
> From: Ryan Schmidt <su...@ryandesign.com>
> Subject: Re: Data caching functionality for subversion
> To: aputagikar@yahoo.com
> Cc: users@subversion.tigris.org
> Date: Wednesday, December 3, 2008, 3:55 PM
> On Dec 2, 2008, at 17:09, Ashish Utagikar wrote:
>
> > Does anybody know whether subversion has any
> data caching functionality. By data caching, I mean the data
> is cached in the cache area so that the the client does not
> have to contact the server every time and it can get its
> data from the cache which usually resides on the same
> machine/file system as the client.
> >
> > Many tools like Design Sync, SOS etc have this
> functionality.
> >
> > Right now it looks like the client has to contact the
> server everytime which might reside on the remote machine
> during the update
>
> It depends on the command. Subversion is designed to be
> used in low-bandwidth situations so yes certainly some
> information is cached in the working copy. For example, the
> .svn directory inside every directory in your working copy
> contains a pristine copy of all the files and their
> properties, so that if you want to see what you've
> changed, "svn diff" does not have to (and does
> not) contact the repository. Same with "svn
> status".
>
> Other commands do contact the repository. "svn
> update" gets changes from the repository and "svn
> commit" sends your changes to the repository so those
> clearly contact the repository. "svn log" gets the
> log from the repository; the log is not cached locally
> presumably because it could be changed after the fact if a
> pre-revprop-change hook is installed to permit that, and
> because the log might be rather large. "svn blame"
> contacts the server because it has to go through the entire
> history of the file and that's not stored locally.
>
> Have you read the book? It might explain more about this.
>
> http://svnbook.org/
>
>
> Note that svk is an alternative to svn, built on top of the
> svn libraries, and that it in fact keeps an entire copy of
> the repository locally, instead of using .svn directories in
> your working copy. I have not used svk, but I believe this
> means you can ask for log information, blame information,
> and even do commits without having access to the master
> repository. svk is compatible with regular svn repositories.
> For more info:
>
> http://svk.bestpractical.com/
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=979222
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].
Re: Data caching functionality for subversion
Posted by Ryan Schmidt <su...@ryandesign.com>.
On Dec 2, 2008, at 17:09, Ashish Utagikar wrote:
> Does anybody know whether subversion has any data caching
> functionality. By data caching, I mean the data is cached in the
> cache area so that the the client does not have to contact the
> server every time and it can get its data from the cache which
> usually resides on the same machine/file system as the client.
>
> Many tools like Design Sync, SOS etc have this functionality.
>
> Right now it looks like the client has to contact the server
> everytime which might reside on the remote machine during the update
It depends on the command. Subversion is designed to be used in low-
bandwidth situations so yes certainly some information is cached in
the working copy. For example, the .svn directory inside every
directory in your working copy contains a pristine copy of all the
files and their properties, so that if you want to see what you've
changed, "svn diff" does not have to (and does not) contact the
repository. Same with "svn status".
Other commands do contact the repository. "svn update" gets changes
from the repository and "svn commit" sends your changes to the
repository so those clearly contact the repository. "svn log" gets
the log from the repository; the log is not cached locally presumably
because it could be changed after the fact if a pre-revprop-change
hook is installed to permit that, and because the log might be rather
large. "svn blame" contacts the server because it has to go through
the entire history of the file and that's not stored locally.
Have you read the book? It might explain more about this.
http://svnbook.org/
Note that svk is an alternative to svn, built on top of the svn
libraries, and that it in fact keeps an entire copy of the repository
locally, instead of using .svn directories in your working copy. I
have not used svk, but I believe this means you can ask for log
information, blame information, and even do commits without having
access to the master repository. svk is compatible with regular svn
repositories. For more info:
http://svk.bestpractical.com/
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=979209
To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].