On Saturday, January 18, 2014 1:51:30 AM UTC-6, Edward K. Ream wrote:
> I am going to completely rewrite phase 2.
Done.
> It's much simpler to ground everything in the actual imported nodes,
*not* a simulated tree that includes organizer nodes.
> Dropping organizer headlines from unls *is* a good idea.
Dropping unls can be done automatically! This is a big collapse in
complexity in phase 1. Not having munge unl's eliminates many potential
errors.
Here is a slightly revised version of the checkin log for one of the merged
components of rev 6596:
QQQ
There has been a breakthrough in phase 1: vc.find_position_for_relative_unl
drops components of unls *as needed* to find a position.
This has many important benefits:
1. It breaks a circular dependency in phase 1: vc.create_organizer_data can
now compute od.anchor without know which unls need dropping.
This dependency was causing all sorts of problems.
2. vc.find_position_for_relative_unl now works without knowing *anything*
about organizer nodes!
3. vc.find_position_for_relative_unl now has the maximal chance of finding
nodes in the outline.
In fact, there is just a slight chance that
vc.find_position_for_relative_unl will find nodes when it shouldn't ;-)
However, I believe that possibility existed in the old way too:
vc.find_position_for_relative_unl ignores child indices: if two siblings
have the same headline, only the suboutline of the first sibling will be
search.
It would be possible to use a recursive, backtracking search to find more
matches, but that's not likely to help much...
QQQ
Edward
--
You received this message because you are subscribed to the Google Groups
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/groups/opt_out.
