Jonathan Gilbert wrote:
> At 04:35 PM 06/07/2005 -0400, I wrote:
> [snip]
>
>>I'm also thinking about ways to efficiently check the checksums of existing
>>files against the repository.
>
>
> After some discussion on the IRC channel and investigation into the code
> that drives the servers, it seems that this actually is not something that
> could be hacked on, albeit inefficiently without a specialized server
> command, but something that is impossible without adding a new command to
> the libsvn_ra interface.
>
> I have cleaned up my patch, and it no longer damages the public API, but it
> still does not avoid downloading files that already exist. I will begin
> poking around at the RA transport providers, and maybe some time in the
> next little while I'll come up with a patch that makes the 'svn takeover'
> command effective as a means to avoid large amounts of data transfer from
> the server.
It seems that there's been a generally positive response to this idea,
and some design discussion, but little review of the actual patch. The
design discussions seem to have died down without a clear resolution as
to whether to go with an "svn takeover" command as provided in this
patch, or to implement rsync functionality to prevent re-downloading.
So I *think* that this patch may be filed as a new issue in the issue
tracker. If anyone feels that I should not do this, let me know.
Otherwise I'll file an issue in the next day or two.
--
Michael W Thelen
It is a mistake to think you can solve any major problems just with
potatoes. -- Douglas Adams
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Jul 27 07:14:42 2005