I was trying to understand the editor functions accessed from configdialog so that I could test those and thought the best way to understand what was really happening was to document and test the functions being called.

Some notes:

1. I haven't done ResetColorizer because that seemed like its own rabbithole.
2. I had to add a windows-only test because os.startfile is a windows only function and wouldn't work on my system. Hope it works.
3. I was getting the timer_event CodeContext errors when running the tests, so I added a line to _close() in editor.py to unbind that event before deleting the text instance and it made the error go away. If it's the right fix, you may want to pull that out into its own PR.

Sorry that this is still massive. I really tried to limit it to what I needed to look at.

Ping, @csabella?

When you're done making the requested changes, leave the comment: I have made the requested changes; please review again.

Whoops, by responding to Tal's comments as part of a new review, my responses got listed twice, once under Tal's comment where they below and then again separately, without the context.

I minimally resolved the merge conflict in the web box so I could proceed. This did a full merge update rather than just updating the one file. There is still the redundancy of global requires gui and class requires gui to be fixed, and Tal's other conflicts. Since I had to do this, I will go ahead and edit everything on this PR.

I stopped the Windows test failures by commenting out the failing lines with an explanation.

The Ubuntu failure is

  File "/home/runner/work/cpython/cpython-ro-srcdir/Lib/idlelib/idle_test/test_editor.py", line 422, in test__extra_help_callback_not_windows
    ehc = partial(w._EditorWindow__extra_help_callback, w)
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
...
AttributeError: Mock object has no attribute '_EditorWindow__extra_help_callback'. Did you mean: '_EditorWindow__recent_file_callback'?

May do the same here.

I think editor.py is more or less commit ready. Tal, review again or not as you wish.

Test_editor is not ready. test_editor passes, but still needs requires('gui') restricted. I want to try Tal's suggested alternative. And also make sure that every test is actually testing. The test I revised was not, as the function supposedly being tested was mocked.

test_mainmenu passes by itself, but test_idle fails with

File "f:\dev\3x\lib\idlelib\idle_test\test_mainmenu.py", line 17, in test_default_keydefs
self.assertGreaterEqual(len(mainmenu.default_keydefs), 50)
AssertionError: 5 not greater than or equal to 50

5 happens to be the length of keydefs used in test_editor, which run first, so the test change is somehow affecting mainmenu. This is probably via changing something in config. test_editor does not import config, but editor does.

Additional test failure: Windows error box titled "Document Start Failure" with message "(X) boom". Test freezes until box is dismissed. I suspect a test of additional help resources needs mocks for both Windows and *nix.

This PR is stale because it has been open for 30 days with no activity.

This PR is stale because it has been open for 30 days with no activity.