[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: RA->get_repos_root() needed

From: <kfogel_at_collab.net>
Date: 2003-10-13 22:21:26 CEST

mark benedetto king <mbk@lowlatency.com> writes:
> The log message receiver gets passed a hash of "changed_paths". The
> keys on this hash are absolute repository-paths (e.g. "/trunk/INSTALL").
> In order to know what key to used to inspect this hash for changes, I need
> to be able to split "http://svn.collab.net/repos/svn/trunk/INSTALL" into its
> repository-URL and path components.
>
> There does not appear to be a straightforward way of doing this. I suspect
> that if there were, issue-960 would have been closed by now.
>
> I think that ra_file and ra_dav can produce their repository-URL reasonably
> easily, but that ra_svn will need a minor protocol change. I'm willing to
> do the work once we reach consensus.

A thought: couldn't you take the initial log message (the youngest
one), find the target file path in the file list, and "lay it over"
the original URL starting from the right edge, using the difference to
discover the non-path portion of the URL?

> Anyway, in the meantime, I have a nasty hack that will work (I'll iterate
> through the keys, taking the shortest one with the longest tail-match on
> my URL).

Sure, whatever works!

So, a question: Is the current blame output formally incorrect under
some circumstances? If it is, we'll need an issue -- which I'm happy
to file, just say the word. We can't enable the feature in 1.0 unless
the output is correct, IMHO. (Nor should we try to solve issue #960
before 1.0, but hopefully this 'blame' problem can be solved another
way -- brute force is fine :-) ).

-Karl

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Oct 13 22:57:34 2003

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.