On 13.01.2014 03:43, Vincent Lefevre wrote:
> On 2014-01-11 22:38:51 +0100, Branko Čibej wrote:
>> On 11.01.2014 10:30, Vincent Lefevre wrote:
>>> Moreover if you want to speed up things, keeping pristines may not
>>> be the right way (in particular if the whole files are kept, not
>>> deltas).
>> FWIW, you can't reasonably store pristines as deltas because you'd have
>> to recalculate them every time someone changes the file in the working
>> copy; and moreover, there's no foolproof way to detect when the WC files
>> are being changed in time to do that. You'd need support from the OS
>> filesystem for that (and I don't mean just file change notifications but
>> copy-on-write per-file snapshots).
> I meant deltas like in the repository (but see below).
When you say "delta" you have to also define "against what". Otherwise
it's just a not very efficient compression algorithm.
>>> Caching the repository or a part of the repository may be better.
>> That's exactly what pristines are.
> Not in a memory efficient way! For one of my working copy, a
> "svn cleanup" reduced the working copy size by about 2 GB, while
> the repository is only 1.5 GB.
Indeed. The point of the whole exercise is to invent an efficient way of
storing (or omitting, when appropriate) pristines. Saying "delta" or
"cache the repository" is just hand-waving and doesn't really contribute
to a solution.
-- Brane
--
Branko Čibej | Director of Subversion
WANdisco // Non-Stop Data
e. brane_at_wandisco.com
Received on 2014-01-13 03:51:56 CET