PR #4188 is now in devel

23 views

Skip to first unread message

Edward K. Ream

unread,

Nov 17, 2024, 8:26:27 AM11/17/24

to leo-editor

PR #4188 removes more cruft from Leo's colorizer.

The PR makes one other change. As explained below, this change should be inconsequential. Nevertheless, please continue testing :-)

Further study

This PR examines the three methods that call jedit.init. I didn't know whether those methods were correct. Now I see that they are.

The PR changes jedit.init so that it no longer clears the state. This change will have no effect because all three callers do (and did) set or clear the state as required. The PR corrects init's docstring to make it less confusing.

A milestone

The previous PRs related to colorizing removed all code that could possibly cause problems with the qsh. The jedit class is the minimal interface with the qsh:

- The qsh calls only recolor.

- recolor is as simple as possible. It contains only one (required!) "if" statement.

Summary

Recent PRs ensure that Leo's colorizer is the minimal interface to the qsh. There appears to be no room for further bugs.

Indeed, I am now convinced that the single "if" state in recolor is correct.

All other "if" statements occur in pattern matchers. They are almost certainly correct. Even if they weren't, they would not affect the qsh in any way.