Re: "svn diff" performance

Ben Collins-Sussman wrote:

>Marcus Comstedt <marcus@mc.pp.se> writes:
>
>
>
>>>We need a way to prevent that resource walk from happening for two of
>>>the three callers to RA->do_switch().
>>>
>>>
>>Sounds like a third variant is needed then. do_update(), do_switch()
>>and do_diff(). What are the other differences (apart from the
>>resource walk) between do_update() and do_switch() that cause diff and
>>merge to not be able to use do_update()?
>>
>>
>
>do_update() assumes you're comparing the same path in two different
>revisions. do_switch() assumes you're comparing two different paths.
>
>The diff and merge command might be comparing anything; the paths
>could be identical, or totally unrelated. So they call do_switch(),
>since that's always guaranteed to work.
>
>Rather than create a whole new do_diff(), maybe we can add a new flag
>to do_switch() that indicates whether we want the vsn-rsc-url data
>sent back to us.
>
>

Yes, and fold do_update into do_switch, if the only difference is the
same-path assumption. Then rename that to do_update again, and use that
everywhere, with appropriate flage.

/me hasn't looked at the code, but pontificates anyway.

-- 
Brane Čibej   <brane_at_xbc.nu>   http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org