On Nov 15, 9:31 am, "Edward K. Ream" <[email protected]> wrote:
> I am going to make this a new thread to emphasize its importance.

Just a few more words about why I have been willing to devote several
weeks of my life to making external unit tests work just like local
unit tests.

1. The new work demonstrates that the nullGui is a completely
equivalent replacement for Leo's qt gui.  This means that the
leoBridge module is completely equivalent to Leo.  That's important,
and may suggest new ways of using the bridge.

2. The new work demonstrates the complete independence of Leo's core
from any particular gui.  True, we don't anticipate a new gui being
necessary, but now the unit tests continually prove that a new gui is
possible.

3. The HighLevelInterface class fully documents the interface between
Leo's core and the gui.  Nothing more is needed, or desirable.  This
interface is *much* simpler and easier to understand than before
because the wretched low-level interface is no more.

As the result of the clarity gained from the HighLevelInterface work,
several pieces of confusing cruft have been eliminated.  The only
cruft worth mentioning is that there are four or five versions of
w.toPythonIndex.  All but one now use common code in g.toPythonIndex.
The other version uses Qt-specific code and is much faster.

To summarize, most of the above are all code-level considerations.
These various collapses in complexity will, as usual, lead to further
improvements.  But the real gains are at the user level:

1. Consistent unit tests, not matter how executed.
2. Consistent operation of the leoBridge with the rest of Leo.

Edward

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/leo-editor?hl=en.

Reply via email to