Add Ctrl+Backspace support by mkitzan · Pull Request #3935 · microsoft/terminal

mkitzan · 2019-12-12T18:52:26Z

Summary of the Pull Request

Changes the Ctrl+Backspace input sequence and how it is processed by InputStateMachineEngine. Now Ctrl+Backspace deletes a whole word at a time on CMD and PS.

References

PR Checklist

Closes Ctrl-backspace does not delete back to the previous wordbreak #755
CLA signed. If not, go over here and sign the CLA
Tests added/passed -> made minor edits to tests
Requires documentation to be updated
I've discussed this with core contributors already. If not checked, I'm ready to accept this work might be rejected in favor of a different grand plan. Issue number where discussion took place: Ctrl-backspace does not delete back to the previous wordbreak #755

Detailed Description of the Pull Request / Additional comments

Changed the input sequence for Ctrl+Backspace to \x1b\x8 so the sequence would pass through _DoControlCharacter. Changed _DoControlCharacter to process \b in a way which forms the correct INPUT_RECORDs to delete whole words.

Validation Steps Performed

Ctrl+Backspace works 🎉

zadjii-msft

Thanks for trying to put together a fix here. However, I don't think this works as the right fix. It unfortunately breaks Ctrl+Bksp in WSL in conhost, by causing conhost to instead emit the VT sequence for Ctrl+Alt+Bksp. Take a look at the following table, composed by running showkey -a in:

gnome-terminal
conhost alone, both with and without the PR applied
conpty by using vtpipeterm to use conhost as the "terminal", both without and with the PR applied

Key Chord	`gnome-terminal`	conhost (before PR)	conhost (with fix)	conpty (before PR)	conpty (with fix)
`Bksp`	`DEL` (0x7f)	`DEL` (0x7f)	`DEL` (0x7f)	`DEL` (0x7f)	`DEL` (0x7f)
`Ctrl+Bksp`	`^H` (0x08)	`^H` (0x08)	`^[^H` (0x1b08)	`^H` (0x08)	`^[^H` (0x1b08)
`h`	`h` (0x68)	`h` (0x68)	`h` (0x68)	`h` (0x68)	`h` (0x68)
`Ctrl+h`	`^H` (0x08)	`^H` (0x08)	`^H` (0x08)	`^H` (0x08)	`^[^H` (0x1b08)
`Del`	`^[[3~`	`^[[3~`	`^[[3~`	`^[[3~`	`^[[3~`
`Ctrl+Del`	`^[[3;5~`	`^[[3;5~`	`^[[3;5~`	`^[[3;5~`	`^[[3;5~`
`Alt+Bksp`	`^[DEL` (0x1b7f)	`^[DEL` (0x1b7f)	`^[DEL` (0x1b7f)
`Ctrl+Alt+Bksp`	`^[^H` (0x1b7f)	`^[^H` (0x1b7f)	`^[^H` (0x1b7f)

Fundamentally, there's a problem where clients reading VT input can't differentiate between Ctrl+Bksp and Ctrl+h because they have the same encoding (^H), but that's not the problem here, so we won't try and fix it. Win32 client apps can differentiate, but not VT (read:WSL) ones.

I think the problem might be that conpty (InputStateMachineEngine) needs to pick one of those two as what it translates ^H to, and we might have picked wrong. Conpty is treating ^H as Ctrl+h, not Ctrl+Bksp. If we simply change InputStateMachineEngine to process a ^H into a Ctrl+Bksp rather than a Ctrl+h, does that fix the bug instead?

This might have some side effects where Win32 client apps running in a conpty session are unable to read a Ctrl+h keypress (and will treat it like Ctrl+Bksp), but I honestly think that's okay.

Below is some more raw output I gathered from `conechokey` while investigating.

Details

Conhost only, Before PR, normal (not VT) input:

# bksp
Down: 1 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: \b (0x8) KeyState: 0x20
Down: 0 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: \b (0x8) KeyState: 0x20

# ctrl+bksp
Down: 1 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x28
Down: 1 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: ⌂  (0x7f) KeyState: 0x28
Down: 0 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: ⌂  (0x7f) KeyState: 0x28
Down: 0 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x20

# h
Down: 1 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: h  (0x68) KeyState: 0x20
Down: 0 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: h  (0x68) KeyState: 0x20

# ctrl+h
Down: 1 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x28
Down: 1 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: \b (0x8) KeyState: 0x28
Down: 0 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: \b (0x8) KeyState: 0x28
Down: 0 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x20

Conpty, Before PR, normal (not VT) input:

# bksp
Down: 1 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: \b (0x8) KeyState: 0x0
Down: 0 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: \b (0x8) KeyState: 0x0

# ctrl+bksp
Down: 1 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x8
Down: 1 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: \b (0x8) KeyState: 0x8
Down: 0 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: \b (0x8) KeyState: 0x8
Down: 0 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x0

# h
Down: 1 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: h  (0x68) KeyState: 0x0
Down: 0 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: h  (0x68) KeyState: 0x0

# ctrl+h
Down: 1 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x8
Down: 1 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: \b (0x8) KeyState: 0x8
Down: 0 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: \b (0x8) KeyState: 0x8
Down: 0 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x0

Conpty, After PR, normal (not VT) input:

Down: 1 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: \b (0x8) KeyState: 0x0
Down: 0 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: \b (0x8) KeyState: 0x0

Down: 1 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x8
Down: 1 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: ⌂  (0x7f) KeyState: 0x8
Down: 0 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: ⌂  (0x7f) KeyState: 0x8
Down: 0 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x0

Down: 1 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: h  (0x68) KeyState: 0x0
Down: 0 Repeat: 1 KeyCode: 0x48 ScanCode: 0x23 Char: h  (0x68) KeyState: 0x0

Down: 1 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x8
Down: 1 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: ⌂  (0x7f) KeyState: 0x8
Down: 0 Repeat: 1 KeyCode: 0x8 ScanCode: 0xe Char: ⌂  (0x7f) KeyState: 0x8
Down: 0 Repeat: 1 KeyCode: 0x11 ScanCode: 0x1d Char: \0 (0x0) KeyState: 0x0

mkitzan · 2019-12-17T17:31:45Z

Thanks for the review @zadjii-msft. That's a lot of good info. I'll work on making InputStateMachine process Ctrl+has Ctrl+Bksp.

mkitzan · 2019-12-17T22:01:40Z

With some new changes making InputStateMachine process ^H as Ctrl+Bksp, CMD and PS will delete whole words. In WSL, ^H does not follow the same path through InputStateMachine as in CMD or PS (apparently). Once I figure out how to send WSL the correct INPUT_RECORD, we should be in business.

mkitzan · 2019-12-17T23:15:36Z

@zadjii-msft WSL standalone doesn't delete whole words when Ctrl+Bksp is pressed, which explains why I was having my problems above. After testing the behavior of ^H with this new commit in showkey -a and conechokey, I believe this solves the problem without messing with the VT sequences like my last version. Thanks for pointing out those command line tools btw, very helpful.

zadjii-msft

Yea this is much better. Thanks for looking into this!

zadjii-msft · 2019-12-18T15:27:58Z

src/terminal/parser/ut_parser/InputEngineTest.cpp

            break;
+        case L'\b': // backspace
+            wch = '\x7f';
+            expectedWch = '\x7f';


nit: this should probably have a break too, even though it's the last case

Makes sense, the latest commit added the break statement

zadjii-msft · 2020-01-09T13:29:28Z

@msftbot merge this in 168 hours

ghost · 2020-01-09T13:29:32Z

Hello @zadjii-msft!

Because you've given me some instructions on how to help merge this pull request, I'll be modifying my merge approach. Here's how I understand your requirements for merging this pull request:

I won't merge this pull request until after the UTC date Thu, 16 Jan 2020 13:29:30 GMT, which is in 7 days

If this doesn't seem right to you, you can tell me to cancel these instructions and use the auto-merge policy that has been configured for this repository. Try telling me "forget everything I just told you".

carlos-zamora

This has been one of my biggest pain points. Thank you!

zadjii-msft · 2020-01-13T23:45:11Z

@msftbot forget everything I just told you

ghost · 2020-01-13T23:45:15Z

Hello @zadjii-msft!

Because you've told me to reset the custom auto-merge settings, I'll use the configured settings for this repository when I'm merging this pull request.

mkitzan · 2020-01-14T00:30:23Z

Glad to have this one merged! Thanks

BalintFarkas · 2020-01-16T06:43:19Z

Thanks for making this @mkitzan , I came here looking for an issue and saw that there's an already merged fix, great!

## Summary of the Pull Request Changes the <kbd>Ctrl+Backspace</kbd> input sequence and how it is processed by `InputStateMachineEngine`. Now <kbd>Ctrl+Backspace</kbd> deletes a whole word at a time (tested on WSL, CMD, and PS).  ## References  ## PR Checklist * [x] Closes #755 * [x] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA * [ ] Tests added/passed -> made minor edits to tests * [ ] Requires documentation to be updated * [x] I've discussed this with core contributors already. If not checked, I'm ready to accept this work might be rejected in favor of a different grand plan. Issue number where discussion took place: #755  ## Detailed Description of the Pull Request / Additional comments Changed the input sequence for <kbd>Ctrl+Backspace</kbd> to `\x1b\x8` so the sequence would pass through `_DoControlCharacter`. Changed `_DoControlCharacter` to process `\b` in a way which forms the correct `INPUT_RECORD`s to delete whole words.  ## Validation Steps Performed <kbd>Ctrl+Backspace</kbd> works 🎉 (cherry picked from commit 2b79bd0)

ghost · 2020-01-27T18:49:01Z

🎉Windows Terminal Preview v0.8.10261.0 has been released which incorporates this pull request.:tada:

Handy links:

Pytal · 2020-01-27T19:39:09Z

Is it possible to get this working with git bash?

DHowett-MSFT · 2020-01-27T19:39:47Z

You’ll want to look into inputrc

wmertens · 2020-04-03T14:01:42Z

I'm on 0.10.781.0 but ctrl-backspace simply seems to send ctrl-h in WSL2 :(

zadjii-msft · 2020-04-03T14:12:46Z

@wmertens You'll want to read #4397

DHowett-MSFT · 2020-04-08T19:59:44Z

🎉 Once again, thanks for the contribution!

This pull request was included in a set of conhost changes that was just
released with Windows Insider Build 19603.

The reverse VT input path in _DoControlCharacter mapped the byte 0x08 to a Ctrl+Backspace key event (VK_BACK with LEFT_CTRL_PRESSED and character 0x7F). This was introduced in PR microsoft#3935 (Jan 2020) to make Ctrl+Backspace delete whole words, which fixed issue microsoft#755. At that time, the forward path (TerminalInput) also sent 0x08 for Ctrl+Backspace, so the mapping was internally consistent, if a bit unusual. In September 2022, PR microsoft#13894 rewrote the forward path to properly implement DECBKM (Backarrow Key Mode). Under the default DECBKM setting, TerminalInput now sends 0x08 for plain Backspace and 0x7F for Ctrl+Backspace. The reverse path was never updated to match, breaking the roundtrip: a Backspace keypress encodes to 0x08, which decodes back as Ctrl+Backspace. The Cygwin project is working on using OpenConsole.exe for its ConPTY support, and a proposed patch series contains an ugly workaround for this bug that bypasses the normal input pipe and injects raw WriteConsoleInput events whenever 0x08 appears in the stream (https://inbox.sourceware.org/cygwin-patches/20260312113923.1528-4-takashi.yano@nifty.ne.jp/). Fixing the bug here avoids the need to accept that workaround. The fix assigns VK_BACK directly and clears writeCtrl, following the same pattern as the existing L'\x1b' case. This avoids the VkKeyScanW roundtrip through _GenerateKeyFromChar, which was the root cause: it reverse-mapped 0x7F to Ctrl+Backspace (since that is the keyboard combination that produces 0x7F), and then writeCtrl added LEFT_CTRL_PRESSED back after modifierState was zeroed. Note that the 0x7F handler (line 221) has the inverse problem: it maps 0x7F to plain Backspace, but the forward path now uses 0x7F for Ctrl+Backspace. That is a separate issue left for a follow-up. Assisted-by: Claude Opus 4.6 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>

In Windows 11, pseudo console has a weird behaviour that the Ctrl-H is translated into Ctrl-Backspace (not Backspace). The analysis in https://inbox.sourceware.org/cygwin-patches/2f8628d2-b79a-95a6-480d-7508375958d5@gmx.de/ shows that the reverse VT input path in conhost's `_DoControlCharacter()` maps the byte 0x08 to a Ctrl+Backspace key event (VK_BACK with LEFT_CTRL_PRESSED and character 0x7F). This was introduced in PR microsoft/terminal#3935 (Jan 2020) to make Ctrl+Backspace delete whole words. In September 2022, PR microsoft/terminal#13894 rewrote the forward path to properly implement DECBKM (Backarrow Key Mode), but the reverse path was never updated to match, breaking the roundtrip. Similary, Backspace (0x7f) is translated into Ctrl-H. Due to this behaviour, `inrec_eq()` in `cons_master_thread()` fails to compare backspace/Ctrl-H events in the input record sequence. This patch is a workaround for the issue that replaces Ctrl-H with backspace (0x7f), which will be translated into Ctrl-H in pseudo console. Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp> Reviewed-by: Johannes Schindelin <Johannes.Schindelin@gmx.de> Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>

mkitzan added 3 commits December 12, 2019 10:44

Adds Ctrl+Backspace support

f3e63d0

Changed key sequence needed to be updated in tests

f31e398

Fix so test INPUT_RECORD is correct

d100fa3

mkitzan marked this pull request as ready for review December 12, 2019 21:39

carlos-zamora requested a review from zadjii-msft December 12, 2019 22:19

mkitzan changed the title ~~Adds Ctrl+Backspace support~~ Add Ctrl+Backspace support Dec 13, 2019

zadjii-msft added Area-Input Related to input processing (key presses, mouse, etc.) Area-TerminalControl Issues pertaining to the terminal control (input, selection, keybindings, mouse interaction, etc.) Product-Terminal The new Windows Terminal. labels Dec 16, 2019

zadjii-msft requested changes Dec 17, 2019

View reviewed changes

ghost added the Needs-Author-Feedback The original author of the issue/PR needs to come back and respond to something label Dec 17, 2019

ghost removed the Needs-Author-Feedback The original author of the issue/PR needs to come back and respond to something label Dec 17, 2019

Changed ^H to ctrl+bksp

2fae701

zadjii-msft approved these changes Dec 18, 2019

View reviewed changes

zadjii-msft added the Needs-Second It's a PR that needs another sign-off label Dec 18, 2019

ghost requested review from DHowett-MSFT, carlos-zamora and miniksa December 18, 2019 15:28

mkitzan added 2 commits December 18, 2019 11:05

Minor changes

ac19743

Merge branch 'master' into Ctrl+Backspace

f98a5f5

ghost requested a review from leonMSFT January 9, 2020 04:42

Fixed issues related to recent InputStateMachineEngine changes

4b1522d

ghost added the AutoMerge Marked for automatic merge by the bot when requirements are met label Jan 9, 2020

zadjii-msft mentioned this pull request Jan 9, 2020

Ctrl+Keys that can't be encoded as VT should still fall through as the unmodified character #3483

Closed

carlos-zamora approved these changes Jan 13, 2020

View reviewed changes

ghost merged commit 2b79bd0 into microsoft:master Jan 13, 2020

mkitzan deleted the Ctrl+Backspace branch January 14, 2020 00:16

ghost mentioned this pull request Jan 27, 2020

Ctrl-backspace does not delete back to the previous wordbreak #755

Closed

This was referenced Jan 29, 2020

Ctrl + h's behavior is differ from command prompt window #4397

Closed

Some keybindings used by PSReadLine aren't getting through #879

Closed

vincentfiestada mentioned this pull request Feb 7, 2020

Ctrl+H not detected properly #4487

Closed

Tyriar mentioned this pull request Jun 9, 2020

use ^H to delete word left in cmd.exe microsoft/vscode#98494

Merged

ghost added the Issue-Bug It either shouldn't be doing this or needs an investigation. label Sep 19, 2020

nico-abram mentioned this pull request Jan 6, 2022

Ctrl+Backspace does not delete whole words in cmd.exe or pwsh.exe wezterm/wezterm#1515

Closed

jonahgraham mentioned this pull request May 11, 2023

TM Terminal: backspace deletes whole word eclipse-cdt/cdt#392

Closed

This pull request was closed.

Conversation

mkitzan commented Dec 12, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary of the Pull Request

References

PR Checklist

Detailed Description of the Pull Request / Additional comments

Validation Steps Performed

Uh oh!

zadjii-msft left a comment

Choose a reason for hiding this comment

Below is some more raw output I gathered from conechokey while investigating.

Uh oh!

mkitzan commented Dec 17, 2019

Uh oh!

mkitzan commented Dec 17, 2019

Uh oh!

mkitzan commented Dec 17, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

zadjii-msft left a comment

Choose a reason for hiding this comment

Uh oh!

zadjii-msft Dec 18, 2019

Choose a reason for hiding this comment

Uh oh!

mkitzan Dec 19, 2019

Choose a reason for hiding this comment

Uh oh!

zadjii-msft commented Jan 9, 2020

Uh oh!

ghost commented Jan 9, 2020

Uh oh!

carlos-zamora left a comment

Choose a reason for hiding this comment

Uh oh!

zadjii-msft commented Jan 13, 2020

Uh oh!

ghost commented Jan 13, 2020

Uh oh!

mkitzan commented Jan 14, 2020

Uh oh!

BalintFarkas commented Jan 16, 2020

Uh oh!

ghost commented Jan 27, 2020

Uh oh!

Pytal commented Jan 27, 2020

Uh oh!

DHowett-MSFT commented Jan 27, 2020

Uh oh!

wmertens commented Apr 3, 2020

Uh oh!

zadjii-msft commented Apr 3, 2020

Uh oh!

DHowett-MSFT commented Apr 8, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

mkitzan commented Dec 12, 2019 •

edited

Loading

Below is some more raw output I gathered from `conechokey` while investigating.

mkitzan commented Dec 17, 2019 •

edited

Loading