...
>
> A couple of comments on the class API:
>
> - _update_needed_children() and _get_needed_keys() sounds like
> good candidates for Graph() or some specialization of it.
True, though they also have side effects like removing texts when there
are no more needed children, etc.
>
> - _update_from_one_parent() the doc string says first parent, why
> not call it _update_from_first_parent() then ? Unless you envision
> some other possible usage...
Sure.
>
> - add_special_text(), hmm, what's that ? The doc string doesn't
> help a lot :-) Does that need to be public ?
It does, as it is used by WorkingTree to add the 'current:' text to be
annotated. (One other benefit of this new code is that 'bzr annotate
NEWS' after a merge doesn't annotate both parents independently... \o/)
>
> jam> 4) Implement a pyrex version of Annotator, to handle some
> jam> inner-loop functionality. (Nicely enough, you can still subclass
> jam> from it.)
>
> It would be nice to defines in pyrex *only* those inner-loops,
> not a requirement to land that patch tough.
>
John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
... needed_ children( ) and _get_needed_keys() sounds like
>
> A couple of comments on the class API:
>
> - _update_
> good candidates for Graph() or some specialization of it.
True, though they also have side effects like removing texts when there
are no more needed children, etc.
> from_one_ parent( ) the doc string says first parent, why from_first_ parent( ) then ? Unless you envision
> - _update_
> not call it _update_
> some other possible usage...
Sure.
>
> - add_special_text(), hmm, what's that ? The doc string doesn't
> help a lot :-) Does that need to be public ?
It does, as it is used by WorkingTree to add the 'current:' text to be
annotated. (One other benefit of this new code is that 'bzr annotate
NEWS' after a merge doesn't annotate both parents independently... \o/)
>
> jam> 4) Implement a pyrex version of Annotator, to handle some
> jam> inner-loop functionality. (Nicely enough, you can still subclass
> jam> from it.)
>
> It would be nice to defines in pyrex *only* those inner-loops,
> not a requirement to land that patch tough.
>
John enigmail. mozdev. org
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://
iEYEARECAAYFAkp ThhgACgkQJdeBCY SNAAN5GACgvRr9A oVrnRJ/ u4Gd2nPEjAil qaerO6ldZRQSOp0 wh
mTcAn3q07M6kiM6
=9LiN
-----END PGP SIGNATURE-----