Serializer_v5._pack_revision and _unpack_revision to become public?

[Robert Collins]

On Sun, 2006-01-15 at 23:33 +0100, David Allouche wrote:
> Hello, I had a very productive week-end, with a couple of nice round
> tuits.
> 
> First I have added calltree (for KCachegrind) support to lsprof
> <http://ddaa.net/blog/python/lsprof-calltree.html>. The patch can be
> also found on my bzr branch <http://ddaa.net/bazaar/bzr.ddaa>.
> 
> Then I used the result to improve the bzrk performance to be usable
> again on trees with a large ancestry. This work can be found on my bzrk
> branch <http://ddaa.net/bazaar/bzrk>.
> 
> Long story short, I ended up creating a cache for the revid->Revision
> dictionary used by bzrk that holds the entire ancestry of the branch
> tip. This cache was initially a pickle, but pickle cannot be used with
> untrusted data, so now it uses an XML format that is a trivial extension
> of the format used for the revision store.
> 
> To read and write this cache, I used the _unpack_revision and
> _pack_revision methods of bzrlib.xml5.Serializer_v5.
> 
> Since there is apparently at least one reasonable (?) use case for using
> these methods from outside bzrlib, I would like them to become public
> methods.

We're deprecating all the XML facilities anyway. So : what does your
extension data look like? Would a black-box storage mechanism that lets
you do much the same thing suffice? 

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/20060116/d6f7eb0c/attachment.pgp