PR #4646 merged into "devel"

[Edward K. Ream]

​PR #4646 is now in Leo's "devel" branch. This PR contains significant improvements to Leo's gui code.

Yes, the PR is too big, but there was no real alternative—all the changes are interrelated.  I fervently hope this will be the last mass update to Leo's gui code.

Rather than obsessing over diffs, I relied on all my testing tools. It was rather like flying on instruments.

The highlight of this PR is a much-improved LeoKeyEvent.__init__ method. It:

- Performs thorough checks on incoming Qt key events.

- Supports enhanced key tracing (--trace=keys).

I'll wait for your comments before merging this code into the "6.8.9" branch. Please continue testing the "devel" branch!

Edward

P.S. Several key-related quirps remain. See #4652. --trace=keys reveals that these issues are "upstream" from  LeoKeyEvent.__init__. That is, the issues happen before the ctor is called. Sometimes Qt calls k.masterKeyHandler without calling the ctor. Sometimes Qt doesn't call either method.

To summarize, the fix for #4652 has little to do with PR #4646. I expect the new PR will be just a few lines long, but finding those lines will be a challenge. Happily, #4652 has a low priority. I have bigger bugs to fix.

EKR

[Edward K. Ream]

On Saturday, May 2, 2026 at 4:16:02 AM UTC-5 Edward K. Ream wrote:

​PR #4646 is now in Leo's "devel" branch. This PR contains significant improvements to Leo's gui code.

Sheesh. There are problems. PR #4655 fixed some of them, but the icon bar somehow got hosed. How did I never notice that?

git bisect has isolated the likely culprit. I'll fix that problem asap.

Edward

[Edward K. Ream]

On Saturday, May 2, 2026 at 5:26:58 AM UTC-5 Edward K. Ream wrote:

> Sheesh. There are problems. PR #4655 fixed some of them, but the icon bar somehow got hosed. How did I never notice that?

> git bisect has isolated the likely culprit. I'll fix that problem asap.

PR #4656 contains the one-line fix. Let the wild rumpus continue!

Edward

[Edward K. Ream]

On Saturday, May 2, 2026 at 5:41:36 AM UTC-5 Edward K. Ream wrote:

> PR #4656 contains the one-line fix. Let the wild rumpus continue!

I should emphasize that only the "devel" branch contains the latest work. That's on purpose: the 6.8.9 branch is for more tested code.

So please keep testing devel. Thanks.

Edward

[Viktor Ransmayr]

Hello Edward,

 I tested the 'devel' branch in a Debian 13 & a Fedora 43 container.

I performed some tests manually w/o finding any issues ...

I also did run the 'full_test_leo' script - and - again only on Fedora the same issues were reported by 'mypy' as last time [ 1 & 2 ].

With kind regards,

Viktor

---

[1] Own F/B related to PR #4649

* https://groups.google.com/g/leo-editor/c/GwuaSGrSJC4/m/OrIru0VDAwAJ

[2] Info related to 'full_test_leo'

* https://groups.google.com/g/leo-editor/c/GwuaSGrSJC4/m/f-m_hz9-AgAJ

[Edward K. Ream]

On Sat, May 2, 2026 at 9:28 AM Viktor Ransmayr <viktor....@gmail.com> wrote:

> I tested the 'devel' branch in a Debian 13 & a Fedora 43 container.
> I performed some tests manually w/o finding any issues ...
> I also did run the 'full_test_leo' script - and - again only on Fedora the same issues were reported by 'mypy' as last time [ 1 & 2 ].

Many thanks for your continued testing. As before, I'm not concerned about pylint and mypy complaints.

Attached is my .pylintrc file. HTH.

Edward