All commit authors signed the Contributor License Agreement.

LGTM. I cannot do a thorough review, but at least confirm, that with this PR I can enter äÄöÖüÜ, etc, again 🚀

LGTM. I cannot do a thorough review, but at least confirm, that with this PR I can enter äÄöÖüÜ, etc, again 🚀

Had some time now: this needs fixes for Linux. Otherwise looks good, thanks @sergey-miryanov!

This breaks Linux, because only one char is inserted in UnixConsole.get_event() [...]

Since this wasn't caught by the tests, we should also probably add a test for that.

I have suggested the def test_push_unicode_character_two_bytes(self): which tests exactly that.

@chris-eibl Please take a look! I would like to keep tests for paste mode to be sure that escape sequences not mixed with input if error occurred. But I'm OK to remove them if you think it is not necessary.

LGTM. Sure, more tests are always better. Though, AFAIU, there is nothing special about paste mode here:

• test_push_unicode_character_two_bytes_in_paste_mode just asserts that a valid "two byte char" within a sequence of "one byte chars" is working as expected.
• likewise, test_push_unicode_character_as_str_in_paste_mode asserts that when an exception during a sequence of feeding bytes via push is caught, we can continue pushing afterwards.

IMHO, both are reasonable, but "paste mode" makes them look too special about pasting?

So I'd rename them and use arbitrary "one byte chars" in there.

Sequence of "one byte chars" is a control command to enable "paste mode". Originally in gh-131878 problem aroise from case where the wrongly passed unicode string puts to the buffer and mixed with following control command that disables paste mode and asserted in those code path. So, I want to check that we don't "corrupt" chars buffer.

Yupp, I know. But as said, there is nothing special about bracketed pasting here. From the perspective of the eventqueue test, these are just arbitrary bytes. Whether bracketed pasting is working correctly, is tested in

If you'd like to showcase that sequences of "one byte chars" interleaved with "multi byte chars" work correctly via "paste mode", then maybe just add a comment?

And small nit: ~ is missing in the escape sequences, see

To me, those two tests have merit, but are too special about "paste mode".

@chris-eibl It seems that I did not correctly assume how paste-mode works (shame on me!). I removed one test because the same code path is covered by test_push_unicode_character_two_bytes. And renamed and simplified a bit test_push_single_chars_and_unicode_character_as_str (not sure the name is ok though).

I don't have a better name - the comment in there clarifies what we intend to test (not much tbh).
Just a small nit on the comment.

Otherwise LGTM. Thanks for bearing with me!

@chris-eibl Thanks for the review and patience!