Fix #1393: avoid dead key check for known unambiguous control characters

[ForNeVeR]

PR Summary

This introduces the fix suggested by @lzybkr in #1393 (comment). I've verified that the wrong behavior is gone after the fix.

PR Checklist

• PR has a meaningful title
  • Use the present tense and imperative mood when describing your changes
• Summarized changes
• Make sure you've added one or more new tests
  • I'm not sure that the changes are easily (if at all) testable; on my machine, many of the tests are failing no matter what and I'm not sure if they should or not
• User-facing changes
  • Not Applicable
  • OR
  • Documentation needed at PowerShell-Docs
    • Doc Issue filed:

Microsoft Reviewers: Open in CodeFlow

[ForNeVeR]

I could use some help with tests.

• Test.en_US_Windows.CharacterSearchBackwardEmacs and Test.en_US_Windows.ShowKeyBindings already fail on my machine even before the changes
• for other three failed tests, I couldn't able to understand what's going on (yet), and I couldn't (yet) find what's been broken in the actual shell

Considering the (very poor!) overall state of the tests on my machine (I have 88 tests failing out of 842), I'm not even sure the test failure is not spurious.

[ForNeVeR]

Regarding other failures: I believe that the test keymap data is filled incorrectly, and that incorrect data was earlier fixed by the code I've turned off in this PR.

For example, let's take a look at the test Test.en_US_Windows.YankLastNth. It fails on the first check, namely, when pressing Ctrl-Alt-y.

On my actual system in runtime, this generates a key event with KeyChar of 25 (or at least that's what I see in the breakpoint set at the line I've changed), which then gets properly converted to y, so the code works in the actual console.

But KeyInfo-en-US-windows.json (the test data file) contains the following:

    {
        "Key":  "Ctrl+Alt+y",
        "KeyChar":  "\u0000",
        "ConsoleKey":  "Y",
        "Modifiers":  "Alt, Control"
    },

The same data (with KeyChar = '\0') was regenerated by Get-KeyInfoData.ps1 script for me, so, something's probably not right here. This test data, in turn, generates not the same ConsoleKeyInfo for tests, and that's why the test fails.

Any ideas on how to proceed?

[daxian-dbw]

I'm using the en-US keyboard layout.
All tests passed without the change in this PR. When applying the change from this PR, I got the same failures as in the CI.

For Ctrl-Alt-y, I got the exact KeyChar, ConsoleKey and Modifiers as shown the above comment.

[daxian-dbw]

The change is a breaking change. At least for key chords like Ctrl+1 and Ctrl+2, they will be interpreted to Ctrl+@ with this change, so any bindings with those chords won't work after this change.

[Den-dp]

Is there something we can do in order to make it a non-breaking change?

[ForNeVeR]

I really could use some help from the team. I'm not able to fix the tests, since they don't work on my computer in the first place.

Probably, we could invent another approach to hotkey handling, but I'm not that savvy with the Windows console API w.r.t. hotkeys, unfortunately.

[terekhov-d]

Any update on this?

[daxian-dbw]

Superseded by #3786