[rfc] [patch] speeding up bzr log with a tree delta iterator

[Robert Collins]

On Thu, 2006-02-16 at 23:38 -0500, Aaron Bentley wrote:
> +        rev1
> +            revno or revid or None.  If it is a revid not in this 


We used to have a lot of apis that were 'revno or revid'... and it took
effort to nuke them. Lets not add such again. -1.

I suggest have the revno api on branch if we really need it, and the
revid 'worker' on repository.

-1 while there is a combined revno-or-revid method.



> +    def delta_iter(self, reviter, reverse=False, delta=False):
> +        """returns an iterator over triples (revno, revid, revdelta).
> +
> +        where revno is the revno of revid if the latter is in the
> revision
> +        history of the branch, None otherwise, and revdelta is None
> if delta
> +        is False, or is the revision delta with respect to the
> preceeding
> +        revid of True.
> +
> +        reviter
> +            an iterator yielding revids.  The oldest revid will not
> be
> +            yielded.

This docstring is either inconsistent or confusion. Also the revno/revid
unclear separation is bad again.  -1 on that part.


Other than that I think its fine.

Theres a bit of a unrelated concern though, which is that 'log' is a
strange mix of UI and engine logic at the moment, which is reflected by
the tests, that are testing correctness by output formatting rather than
object value. Not something to worry about for this patch though.

Rob


 
-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060223/691cff29/attachment.pgp