starred/terminal
1
0
Fork 0
mirror of https://github.com/microsoft/terminal.git synced 2026-02-04 05:35:20 +00:00
Code Issues 19.0k Packages Projects Releases 20 Wiki Activity

Incorrect Shift / Meta / Control bits set in xterm mouse mode #11443

New Issue
Closed
opened 2026-01-31 02:47:48 +00:00 by claunia · 11 comments
claunia commented 2026-01-31 02:47:48 +00:00
Owner
Copy Link

Originally created by @chaosemer on GitHub (Nov 16, 2020).

Originally assigned to: @DHowett on GitHub.

Environment

  • Windows build number: Microsoft Windows [Version 10.0.19042.630]
  • Windows Terminal version (if applicable): 1.4.3141.0
    • (This also applies to the legacy Windows Command Prompt app.)
  • Debian version (cat /etc/debian_version): 10.6

Steps to reproduce

  1. Run debian.
  2. Enable xterm mouse tracking mode: 
    $ echo -e '\e[?1000h'
  3. Click on the upper right-most character with the left mouse button. Try doing this with a mix of Ctrl, Alt, and/or Shift held down.

Expected behavior

Per https://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h2-Mouse-Tracking, we should get back six characters on a mouse button down event followed by an additional six characters on a mouse button up event. The first three characters (ESC [ M) will not be displayed, only the last three characters per event get displayed - this is not a bug, it is just important to understand.

  • A mouse button 1 down event with no modifiers held should display the following three characters: Space ! !
  • A mouse button 1 down event with control held should display the following three characters: 0 ! !
  • A mouse button 1 down event with alt held should be treated as meta and display the following three characters: ( ! !
  • A mouse button 1 down event with shift held should display the following three characters: $ ! !
  • A mouse button 1 down event with control and alt held should display the following three characters: 8 ! !

Actual behavior

  • With control held, you get ( ! ! as if meta is held.
  • With alt held, you get Space ! !, as if nothing is held.
  • With shift held, you select characters (This is likely not a bug, just an override that I did not disable.)
  • With control and alt held, you get ( ! !, as if only meta is held.
Originally created by @chaosemer on GitHub (Nov 16, 2020). Originally assigned to: @DHowett on GitHub. # Environment * **Windows build number:** Microsoft Windows [Version 10.0.19042.630] * **Windows Terminal version (if applicable):** 1.4.3141.0 * (This also applies to the legacy Windows Command Prompt app.) * **Debian version (cat /etc/debian_version):** 10.6 # Steps to reproduce 1. Run debian. 2. Enable xterm mouse tracking mode: ```bash $ echo -e '\e[?1000h' ```` 3. Click on the upper right-most character with the left mouse button. Try doing this with a mix of Ctrl, Alt, and/or Shift held down. # Expected behavior Per https://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h2-Mouse-Tracking, we should get back six characters on a mouse button down event followed by an additional six characters on a mouse button up event. The first three characters (`ESC` `[` `M`) will not be displayed, only the last three characters per event get displayed - this is not a bug, it is just important to understand. * A mouse button 1 down event with no modifiers held should display the following three characters: `Space` `!` `!` * A mouse button 1 down event with control held should display the following three characters: `0` `!` `!` * A mouse button 1 down event with alt held should be treated as meta and display the following three characters: `(` `!` `!` * A mouse button 1 down event with shift held should display the following three characters: `$` `!` `!` * A mouse button 1 down event with control and alt held should display the following three characters: `8` `!` `!` # Actual behavior * With control held, you get `(` `!` `!` as if meta is held. * With alt held, you get `Space` `!` `!`, as if nothing is held. * With shift held, you select characters (This is likely not a bug, just an override that I did not disable.) * With control and alt held, you get `(` `!` `!`, as if only meta is held.
claunia commented 2026-01-31 02:47:50 +00:00
Author
Owner
Copy Link

@DHowett commented on GitHub (Nov 16, 2020):

Whoa. You're right. How did we miss this?

@DHowett commented on GitHub (Nov 16, 2020): _Whoa._ You're right. How did we miss this?
claunia commented 2026-01-31 02:47:50 +00:00
Author
Owner
Copy Link

@DHowett commented on GitHub (Nov 16, 2020):

6115f8db82/src/terminal/input/mouseInput.cpp (L134-L137)

6115f8db82/src/terminal/input/mouseInput.cpp (L191-L196)

That's not great.

@DHowett commented on GitHub (Nov 16, 2020): https://github.com/microsoft/terminal/blob/6115f8db82a1fdf1a1efd415426b950327288cb2/src/terminal/input/mouseInput.cpp#L134-L137 https://github.com/microsoft/terminal/blob/6115f8db82a1fdf1a1efd415426b950327288cb2/src/terminal/input/mouseInput.cpp#L191-L196 That's not great.
claunia commented 2026-01-31 02:47:50 +00:00
Author
Owner
Copy Link

@chaosemer commented on GitHub (Nov 17, 2020):

Yup. The control fix is easy enough. It's just wrong in those comments and code. Control should add 0x10 and meta should add 0x08.

@chaosemer commented on GitHub (Nov 17, 2020): Yup. The control fix is easy enough. It's just wrong in those comments and code. Control should add 0x10 and meta should add 0x08.
claunia commented 2026-01-31 02:47:50 +00:00
Author
Owner
Copy Link

@DHowett commented on GitHub (Nov 17, 2020):

I'm alarmed that we missed this for every windows release after (and including) 14393! It was introduced broken back then 😄 and hasn't been caught, or fixed, since.

I guess I have to recommend SGR encoding for now. ☹️

EDIT: I rescind the frowning face above since SGR encoding is superior in every way.

@DHowett commented on GitHub (Nov 17, 2020): I'm alarmed that we missed this for every windows release after (and including) 14393! It was introduced broken back then :smile: and hasn't been caught, or fixed, since. I guess I have to recommend SGR encoding for now. ☹️ EDIT: I rescind the frowning face above since SGR encoding is superior in every way.
claunia commented 2026-01-31 02:47:50 +00:00
Author
Owner
Copy Link

@jdebp commented on GitHub (Nov 17, 2020):

The ECMA-48 conformant format of DEC private mode 1006 is indeed the better format of mouse report, and isn't just a "for now" recommendation, really. It's a "from now on" recommendation.

Yes, a terminal emulator should get the mouse report formats of DEC private modes 9, 1005, and 1015 right if it supports them at all, but they are part of an evolutionary process that the world of XTerm and RXVT went through back around 2010 to 2012, with DEC private mode 1006, and the conseqential private control sequence CSI < M , being the upshot.

Applications that use terminfo will nowadays use DEC private mode 1006 on a wide variety of terminal types, from genuine XTerm through iTerm2 and Konsole to libVTE terminal emulators, and won't in fact use DEC private modes 1005 or 1015 on any terminal type, as no entry in a modern terminfo database employs them. Terminfo applications will furthermore only use the old X10/X11 report format with just three remaining terminal types: screen, terminology, and iTerm.

Interestingly, the current terminfo database doesn't yet list Windows Terminal (ms-terminal) as having an XM capability.

When that is updated, then tput XM 1 should suffice for setting mouse reports on from the command line, rather than emitting control sequences longhand, as should setterm -7 --xterm-mouse-reports all with my portable setterm tool.

@jdebp commented on GitHub (Nov 17, 2020): The ECMA-48 conformant format of DEC private mode 1006 is indeed the better format of mouse report, and isn't just a "for now" recommendation, really. It's a "from now on" recommendation. Yes, a terminal emulator should get the mouse report formats of DEC private modes 9, 1005, and 1015 _right_ if it supports them at all, but they are part of an evolutionary process that the world of XTerm and RXVT went through back around 2010 to 2012, with DEC private mode 1006, and the conseqential private control sequence CSI < M , being the upshot. Applications that use terminfo will nowadays use DEC private mode 1006 on a wide variety of terminal types, from genuine XTerm through iTerm2 and Konsole to libVTE terminal emulators, and won't in fact use DEC private modes 1005 or 1015 on any terminal type, as no entry in a modern terminfo database employs them. Terminfo applications will furthermore only use the old X10/X11 report format with just three remaining terminal types: screen, terminology, and iTerm. Interestingly, the current terminfo database doesn't yet list Windows Terminal ([`ms-terminal`](https://invisible-island.net/ncurses/terminfo.src.html#tic-ms-terminal)) as having an `XM` capability. When that is updated, then `tput XM 1` should suffice for setting mouse reports on from the command line, rather than emitting control sequences longhand, as should `setterm -7 --xterm-mouse-reports all` with [my portable `setterm` tool](http://jdebp.uk./Softwares/nosh/guide/commands/setterm.xml).
claunia commented 2026-01-31 02:47:51 +00:00
Author
Owner
Copy Link

@DHowett commented on GitHub (Nov 17, 2020):

Alas, that terminfo entry was built when Terminal 0.2 shipped without considering the things supported by the underlying console host. We’ve had XM (apparently with the caveats mentioned here) for three or so years 😄
We’ve been in touch with the maintainer on and off since then. We’re closer to xterm+indirect with a few capabilities blanked than to the ms-terminal entry that has been shipping.

@DHowett commented on GitHub (Nov 17, 2020): Alas, that terminfo entry was built when Terminal 0.2 shipped without considering the things supported by the underlying console host. We’ve had XM (apparently with the caveats mentioned here) for three or so years :smile: We’ve been in touch with the maintainer on and off since then. We’re closer to xterm+indirect with a few capabilities blanked than to the ms-terminal entry that has been shipping.
claunia commented 2026-01-31 02:47:51 +00:00
Author
Owner
Copy Link

@jdebp commented on GitHub (Nov 17, 2020):

Unfortunately, I couldn't get mouse reporting to work at all here with 1.4.3141.0, but I suspect that ssh being in the mix in my testing of it is partly the cause of that.

terminfo is a side issue here. So see #8303. Back to mouseInput.cpp. ☺

@jdebp commented on GitHub (Nov 17, 2020): Unfortunately, I couldn't get mouse reporting to work at all here with 1.4.3141.0, but I suspect that `ssh` being in the mix in my testing of it is partly the cause of that. terminfo is a side issue here. So see #8303. Back to `mouseInput.cpp`. &#9786;
claunia commented 2026-01-31 02:47:51 +00:00
Author
Owner
Copy Link

@DHowett commented on GitHub (Nov 17, 2020):

Ah yes, the SSH client shipped inbox (calls itself 7.7) synthesizes its own VT input from console event records. By 8.1+, we’d finally convinced them to let us handle it with ENABLE_VIRTUAL_TERMINAL_INPUT 😁

@DHowett commented on GitHub (Nov 17, 2020): Ah yes, the SSH client shipped inbox (calls itself 7.7) synthesizes its own VT input from console event records. By 8.1+, we’d finally convinced them to let us handle it with `ENABLE_VIRTUAL_TERMINAL_INPUT` 😁
claunia commented 2026-01-31 02:47:51 +00:00
Author
Owner
Copy Link

@DHowett commented on GitHub (Nov 18, 2020):

Looking inside, I'm deeply horrified by what I see.

  • We made the same mistake for X10 and SGR encoding - control comes out as meta.
  • Terminal (uniquely) treats left/right control, left/right shift and left/right alt differently
    • We pass those control key states through indiscriminately as MK mouse key values
    • This means that left control is meta (until we fix it) and right control is shift (if we re-enable support for MK_SHIFT)

It's almost enough to suggest that perhaps no version of Terminal has ever really supported mouse mode, and conhost has only barely done so until you held down a modifier.

@DHowett commented on GitHub (Nov 18, 2020): Looking inside, I'm deeply horrified by what I see. * We made the same mistake for X10 and SGR encoding - control comes out as meta. * Terminal (uniquely) treats left/right control, left/right shift and left/right alt differently * We pass those control key states through indiscriminately as `MK` mouse key values * This means that left control is meta (until we fix it) and right control is shift (if we re-enable support for `MK_SHIFT`) It's almost enough to suggest that perhaps no version of Terminal has ever _really_ supported mouse mode, and conhost has only barely done so until you held down a modifier.
claunia commented 2026-01-31 02:47:51 +00:00
Author
Owner
Copy Link

@ghost commented on GitHub (Jan 28, 2021):

:tada:This issue was addressed in #8379, which has now been successfully released as Windows Terminal v1.5.10271.0.🎉

Handy links:

@ghost commented on GitHub (Jan 28, 2021): :tada:This issue was addressed in #8379, which has now been successfully released as `Windows Terminal v1.5.10271.0`.:tada: Handy links: * [Release Notes](https://github.com/microsoft/terminal/releases/tag/v1.5.10271.0) * [Store Download](https://www.microsoft.com/store/apps/9n8g5rfz9xk3?cid=storebadge&ocid=badge)
claunia commented 2026-01-31 02:47:51 +00:00
Author
Owner
Copy Link

@ghost commented on GitHub (Jan 28, 2021):

:tada:This issue was addressed in #8379, which has now been successfully released as Windows Terminal Preview v1.6.10272.0.🎉

Handy links:

@ghost commented on GitHub (Jan 28, 2021): :tada:This issue was addressed in #8379, which has now been successfully released as `Windows Terminal Preview v1.6.10272.0`.:tada: Handy links: * [Release Notes](https://github.com/microsoft/terminal/releases/tag/v1.6.10272.0) * [Store Download](https://www.microsoft.com/store/apps/9n8g5rfz9xk3?cid=storebadge&ocid=badge)
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
⛺ Reserved
A11yCO
A11yMAS
A11ySev1
A11ySev2
A11ySev3
A11yTTValidated
A11yUsable
A11yVoiceAccess
A11yWCAG
Area-Accessibility
Area-AtlasEngine
Area-AzureShell
Area-Build
Area-Build
Area-Chat
Area-CmdPal
Area-CodeHealth
Area-Commandline
Area-CookedRead
Area-DefApp
Area-Extensibility
Area-Fonts
Area-GroupPolicy
Area-i18n
Area-Input
Area-Interaction
Area-Interop
Area-Localization
Area-Output
Area-Performance
Area-Portable
Area-Quality
Area-Remoting
Area-Rendering
Area-Schema
Area-Server
Area-Settings
Area-SettingsUI
Area-ShellExtension
Area-ShellExtension
Area-ShellExtension
Area-Suggestions
Area-Suggestions
Area-TerminalConnection
Area-TerminalControl
Area-Theming
Area-UserInterface
Area-VT
Area-Windowing
Area-WPFControl
AutoMerge
Blocking-Ingestion
Culprit-Centennial
Culprit-WinUI
Disability-All
Disability-Blind
Disability-LowVision
Disability-Mobility
External-Blocked-WinUI3
Fixed
Gathering-Data
good first issue
HCL-E+D
HCL-WindowsTerminal
Help Wanted
Impact-Compatibility
Impact-Compliance
Impact-Correctness
Impact-Visual
In-PR
InclusionBacklog
InclusionBacklog-Windows TerminalWin32
InclusionCommitted-202206
Issue-Bug
Issue-Docs
Issue-Feature
Issue-Feature
Issue-Question
Issue-Samples
Issue-Scenario
Issue-Task
Needs-Attention
Needs-Author-Feedback
Needs-Bisect
Needs-Discussion
Needs-Repro
Needs-Tag-Fix
Needs-Tag-Fix
Needs-Triage
No-Recent-Activity
Priority-0
Priority-1
Priority-2
Priority-3
Product-Cmd.exe
Product-Colortool
Product-Colortool
Product-Colortool
Product-Conhost
Product-Conpty
Product-Meta
Product-Powershell
Product-Terminal
Product-WSL
pull-request
Mirrored from GitHub Pull Request
 Resolution-Answered
Resolution-By-Design
Resolution-Duplicate
Resolution-External
Resolution-Fix-Available
Resolution-Fix-Committed
Resolution-No-Repro
Resolution-Won't-Fix
Severity-Blocking
Severity-Crash
Severity-DataLoss
spam
this-will-be-a-breaking-change
Tracking-External
WindowsTerminal_Win32
Work-Item
zAskModeBug
zInbox-Bug
No Label Area-VT Impact-Correctness Issue-Bug Priority-2 Product-Conhost Product-Terminal Resolution-Fix-Committed
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
claunia
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: starred/terminal#11443
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?