Tickets #4872, #4873 and #4874: fix 'insert control code' function

[peripherium]

The previous implementation used ascii_alpha_to_cntrl(), which assumed
alphabetic input and did not properly validate key codes returned by the
terminal libraries. The function has been renamed to keycode_to_cntrl()
and extended to correctly validate and convert key codes.

Changes:

• correctly handle KEY_M_CTRL combinations (e.g. Ctrl+J).

  • Fixes mcedit: Display corruption (missing header) after inserting literal ^J #4872.

• ignore special keys returned by get_key_code() (e.g. function or cursor keys).

  • Fixes mcedit: Insert literal, followed by special key acts weirdly #4873.

• extend control mappings for @ [ \ ] ^ _.

  • Fixes mcedit: Insert literal should accept plain non-letters #4874.

• return -1 for invalid input and ignore it at the call sites

The function is used both in the Insert literal dialog in mcedit and in the mc command line.

Checklist

• I have referenced the issue(s) resolved by this PR (if any)
• I have signed-off my contribution with git commit --amend -s
• Lint and unit tests pass locally with my changes (make indent && make check)
• I have added tests that prove my fix is effective or that my feature works
• I have added the necessary documentation (if appropriate)

[zyv]

Since you are fixing keycode_to_cntrl, could you please add a test for it? Thank you!

[zyv]

@egmontkob adding you, since this PR fixes 3 issues that you have reported...

[peripherium]

I initially didn't add a test for this function because the previous implementation didn't have one, so I assumed that adding a test was not required.

In the meantime, I have taken a closer look at the test suite and added the requested test here.

If this test is not sufficient or needs adjustments, please let me know.

[mc-worker]

Formatting issues:
https://github.com/MidnightCommander/mc/actions/runs/22849421605/job/66290234418

[egmontkob]

The commit message is duplicated.

[egmontkob]

Tested now, and it indeed fixes those three bugs, thanks!

[peripherium]

I included the lowercase characters because I thought it would be user-friendly: even if shift wasn't pressed during input, the corresponding control code can be entered via the dialog. However, this is a bit inconsistent with the last inserted condition, according to which ch == '?' returns 0x7f. Should I remove the conditions with the lowercase characters from the code?

[egmontkob]

Lowercase letters are absolutely fine, in fact, expected; the user probably won't bother pressing Shift.

What's unnecessary IMHO is support the characters "surrounding" the lowercase letters in the ASCII table: ` { | } ~. Here it's absolutely reasonable that the user will instead press @ [ \ ] ^ (or _ which didn't have a counterpart).

[peripherium]

Originally I allowed characters from ´ to ~ because I assumed someone looking at the ASCII table might wonder why those printable characters were excluded.

However, if this is not intuitive and could instead confuse future readers, then it doesn't belong in the code. So I removed it and restricted the range to a-z.

[peripherium]

The commit message is duplicated.

Sorry about the duplicated commit message - that must have slipped in during the amend commits while I was updating the patch. I've fixed the commit message now.

Please let me know if there is anything else in the patch that should be improved.

[egmontkob]

Just very quickly skimming through it this time, it looks good, thank you!

(Super minor nitpicking: you add an extra newline to the end of tests/lib/Makefile.am which already has enough (some will say that already has more than enough.))

Let's wait for Andrew's review as well.

[mc-worker]

Please begin the commit message with following line:

Ticket #4872, #4873, #4874: fix 'insert control code' function.

[mc-worker]

1. Not all test are passed. util__keycode_to_cntrl.log is attached.
   util__keycode_to_cntrl.log

2. In editor, I use a paste of \t via "Insert literal" dialog (ctrl-q) very often. It doesn't work anymore.

[peripherium]

I completely forgot to adjust the test after changing the keycode_to_cntrl() function. Thanks for pointing that out. I've already implemented all the other changes (especially in Makefile.am).

[zyv]

I'm happy with the code now, and I’m very happy about the test. Thank you very much! How about the others?

[egmontkob]

Looked good to me the last time, won't check it again :P Thanks!