On Monday, May 15, 2017 at 10:28:49 AM UTC-5, Edward K. Ream wrote:
*> tl;dr:* We need an interface *class *that delivers (wrappers to) Leo's
position and/or vnode classes instead of some (as yet unknown) npyscreen
*array*.
...
> But oh my, the complications involved in the switcheroo itself are
mind-boggling. I have been studying the code for days, trying to understand
every little wrinkle.
It's never a good idea to "predict" difficulties, or to dwell on
complications. My previous strategy of studying *all *the sources was
inefficient. There is no need to do that! Instead, tracing the *name* of a
method reveals the method (in whatever superclass!) that will actually be
called.
For example, LeoMLTree.delete_line isn't working. The line *appears* to be
deleted from the screen, but collapsing and expanding the deleted node's
parent shows the deleted node again. This bug is directly related to the
_my_widgets and the switcheroo, so I want to fix it asap.
The LeoMLTree.delete_line sources give no indications of the class to which
self.display refers. No problem! Just add g.trace(self.display). The
result will be:
<bound method Widget.display of <leo.plugins.cursesGui2.LeoMLTree
object at 0x086E7B10>>
So now we know that *Widget*.display is being executed, not the display
method of any other class. This trick saves a lot of sleuthing. No need for
a debugger, which is how I previously discovered the actual method being
called.
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 https://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/d/optout.
