starred/terminal
1
0
Fork 0
mirror of https://github.com/microsoft/terminal.git synced 2026-07-08 18:16:28 +00:00
Code Issues 19.0k Packages Projects Releases 20 Wiki Activity

Please finish support for OSC 4 and friends #467

New Issue
Closed
opened 2026-01-30 21:53:00 +00:00 by claunia · 16 comments
claunia commented 2026-01-30 21:53:00 +00:00
Owner
Copy Link

Originally created by @Jaykul on GitHub (Dec 6, 2018).

Originally assigned to: @zadjii-msft on GitHub.

I was really happy to see support for OSC 4 to set the 16 palette colors (and I recently helped a friend add it to the docs, the way it is now) ...

However, I have some problems. I'm hoping you can address them all as one request, but I'm happy to split them out. First, as background the xterm control sequences docs. Obviously these are not sacred, but they do represent a sort-of standard in that other terminals and tmux have implemented these same sequences the same way...

  1. OSC 4 should support multiple colors. XTerm and tmux (and others) support multiple pairs of index;spec without having to repeat the ESC]4; part, so you can do: ESC]4;0;rgb:00/00/00;1;rgb:00/00/70;...\\ and list out the whole set of colors as one giant escape sequence.
  2. It doesn't allow setting the extended xterm 88 or 256 color palette (see #313).
  3. It doesn't support querying. I should be able to put a "?" in place of the rgb spec and get back a control sequence (which is supposed to be the exact same format), such that it could be used to set the color to the same RGB value it is already. This would allow me to check if I'm dealing with PowerShell blue instead of magenta by writing ESC]4;5;?\\ and parsing the output

Additionally ... without support for OSC 10 and 11 I cannot change the default foreground and background colors, so there's no way to fix (even temporarily) the fact that PowerShell is using 5 as it's default background color. NOTE that these should also support querying with a "?"

AND EVEN WORSE ...

We really also need support for OSC 104 (and 110 and 111) which would reset the colors changed by OSC 4, 10, and 11 to their defaults. Note that if done right, they support resetting individual colors, but as a first draft, I'd settle for the parameterless reset of the entire table -- and for my own purposes, if you implement 110 and 111 as meaning force WHITE text on a BLACK background, without implementing 10 and 11, I wouldn't cry 😉

Originally created by @Jaykul on GitHub (Dec 6, 2018). Originally assigned to: @zadjii-msft on GitHub. I was really happy to see support for OSC 4 to **set** the 16 palette colors (and I recently helped a friend add it to the docs, _the way it is now_) ... However, I have some problems. I'm hoping you can address them all as one request, but I'm happy to split them out. First, as background the [xterm control sequences docs](http://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h2-Operating-System-Commands). Obviously these are not sacred, but they _do_ represent a sort-of standard in that _other_ terminals and tmux have implemented these same sequences the same way... 1. OSC 4 should support multiple colors. XTerm and tmux (and others) support multiple _pairs_ of `index;spec` without having to repeat the `ESC]4;` part, so you can do: `ESC]4;0;rgb:00/00/00;1;rgb:00/00/70;...\\` and list out the whole set of colors as one giant escape sequence. 2. It doesn't allow setting the extended xterm 88 or 256 color palette (see #313). 3. It doesn't support querying. I should be able to put a "?" in place of the rgb spec and get back a control sequence (which is supposed to be the exact same format), such that it could be used to set the color to the same RGB value it is already. This would allow me to _check_ if I'm dealing with PowerShell blue instead of magenta by writing `ESC]4;5;?\\` and parsing the output Additionally ... without support for OSC 10 and 11 I cannot change the default foreground and background colors, so there's no way to _fix_ (even temporarily) the fact that PowerShell is using 5 as it's default background color. NOTE that these should _also_ support querying with a "?" ## AND EVEN WORSE ... We really also need support for OSC 104 (and 110 and 111) which would **reset** the colors changed by OSC 4, 10, and 11 to their defaults. Note that if done right, they support resetting individual colors, but as a first draft, I'd settle for the parameterless reset of the entire table -- and for my own purposes, if you implement 110 and 111 as meaning force WHITE text on a BLACK background, _without_ implementing 10 and 11, I wouldn't cry 😉
claunia added the Issue-FeatureProduct-ConhostResolution-Fix-AvailableArea-VT labels 2026-01-30 21:53:00 +00:00
claunia commented 2026-01-30 21:53:03 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Dec 7, 2018):

Okay, this is a great writeup.

However, there's a LOT of work to be done here - more than I can likely justify getting into the next (19H1) release of Windows. That being said, I have filed a pile of internal tasks on me to make sure these scenarios are all covered.

As you called out, #313 is fortunately on my tasklist for this month, so hopefully that one should be out soontm.

This is deliverable MSFT:19859141

@zadjii-msft commented on GitHub (Dec 7, 2018): Okay, this is a great writeup. However, there's a LOT of work to be done here - more than I can likely justify getting into the next (19H1) release of Windows. That being said, I have filed a pile of internal tasks on me to make sure these scenarios are all covered. As you called out, #313 is fortunately on my tasklist for this month, so hopefully that one should be out soon<sup>tm</sup>. This is deliverable MSFT:19859141
claunia commented 2026-01-30 21:53:03 +00:00
Author
Owner
Copy Link

@Jaykul commented on GitHub (Dec 7, 2018):

For what it's worth, @zadjii-msft I would say OSC 110 and 111 (if implemented such that they return it to white on black) would be way more useful than any of the rest of it.

Honestly, the way it is now (or with #313 fixed) it's not very useful. Without knowing/controlling which color is the actual (default) background and foreground, setting the color palette is almost pointless -- did you see that screenshot on #81 with the pink background and bright yellow text?

@Jaykul commented on GitHub (Dec 7, 2018): For what it's worth, @zadjii-msft I would say OSC 110 and 111 (if implemented such that they return it to white on black) would be _way_ more useful than **any** of the rest of it. Honestly, the way it is now (or with #313 fixed) it's not very useful. Without knowing/controlling which color is the _actual_ (default) background and foreground, setting the color palette is almost pointless -- did you see [that screenshot](https://github.com/Microsoft/console/issues/81#issuecomment-413887664) on #81 with the pink background and _bright yellow_ text?
claunia commented 2026-01-30 21:53:03 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Dec 7, 2018):

Ah, so I'd say the main problem you're having is that powershell has hard-coded it's default colors to be yellow on magenta, then re-defined what yellow and magenta are (to a light grey and Powershell Blue). We've tried to get them to change this in the past, but it's never happened.

Currently, o of the big differences between conhost and other terminals is that we don't support the notion of a default color that's not a part of the color table.

I'd have to investigate, but presumably, using those OSC's to reset the "default color" would revert the default color to whatever conhost had when it was launched, but it wouldn't change which indices were set as the defaults.

Also stay tuned in this area, I think we'll have something bubbling up the source tree that I can't really talk about quite yet :)

@zadjii-msft commented on GitHub (Dec 7, 2018): Ah, so I'd say the main problem you're having is that powershell has hard-coded it's default colors to be yellow on magenta, then re-defined what yellow and magenta are (to a light grey and Powershell Blue). We've tried to get them to change this in the past, but it's never happened. Currently, o of the big differences between conhost and other terminals is that we don't support the notion of a default color that's not a part of the color table. I'd have to investigate, but presumably, using those OSC's to reset the "default color" would revert the default color to whatever conhost had when it was launched, but it wouldn't change which indices were set as the defaults. Also stay tuned in this area, I think we'll have something bubbling up the source tree that I can't really talk about quite yet :)
claunia commented 2026-01-30 21:53:04 +00:00
Author
Owner
Copy Link

@oising commented on GitHub (Dec 7, 2018):

  1. OSC 4 should support multiple colors. XTerm and tmux (and others) support multiple pairs of index;spec without having to repeat the ESC]4; part, so you can do: ESC]4;0;rgb:00/00/00;1;rgb:00/00/70;...\\ and list out the whole set of colors as one giant escape sequence.

There is one thing though I'd be careful about - it's that (CSI) control sequences may have multiple parameters, but to be strictly compatible there's a limit of sixteen parameters as anything beyond that is silently ignored. I'm not sure if the same limits apply to OSC sequences, but it's likely. But if tmux/xterm/screen are doing it, I guess it's probably okay?

Source: https://vt100.net/emu/dec_ansi_parser

@oising commented on GitHub (Dec 7, 2018): > 1. OSC 4 should support multiple colors. XTerm and tmux (and others) support multiple _pairs_ of `index;spec` without having to repeat the `ESC]4;` part, so you can do: `ESC]4;0;rgb:00/00/00;1;rgb:00/00/70;...\\` and list out the whole set of colors as one giant escape sequence. There is one thing though I'd be careful about - it's that (CSI) control sequences may have multiple parameters, but to be strictly compatible there's a limit of sixteen parameters as anything beyond that is silently ignored. I'm not sure if the same limits apply to OSC sequences, but it's likely. But if tmux/xterm/screen are doing it, I guess it's probably okay? Source: https://vt100.net/emu/dec_ansi_parser
claunia commented 2026-01-30 21:53:04 +00:00
Author
Owner
Copy Link

@Jaykul commented on GitHub (Dec 7, 2018):

I don't think that applies @oising. CSI parameters are also limited to be numbers, because the thing is building a buffer of parameters to pass to the command. For OSC, there's no buffer: OSC strings have to be terminated by ST, CAN, SUB or ESC -- any other characters are supposedly just "passed" from the control string to the handler as they arrive...

@zadjii-msft I'll retract what I said earlier, if that's how it is. In that case, OSC 110 and 111 are not what I want. I obviously need OSC 10 and 11 instead. 😬 I guess that really, one way or another, I need to be able to force the background to a known value.

I have been assuming that the windows console wouldn't suddenly learn how to have a different default bg/fg and would be limited to one of the 16 palette values ... I'm used to that by now. But I was thinking the defaults would be defaults in the global sense, or at least in the sense that's used in the window's menu...

image

So yeah. Anyway. From my perspective, I don't really care if the background is black, blue, magenta -- but I do need know for certain which color index the background is.

The goal is to avoid that awkward situation where I write text in the background color, when I thought I was writing highly visible pink or white on black ...

And yes, the number one offender is the PowerShell blue background -- but people can (and do) change their defaults. I have had several co-workers with vision impairments and they all tend to force their consoles to black on a white background.

@Jaykul commented on GitHub (Dec 7, 2018): I don't think that applies @oising. CSI parameters are also limited to be numbers, because the thing is building a buffer of parameters to pass to the command. For OSC, there's no buffer: OSC strings have to be terminated by ST, CAN, SUB or ESC -- any other characters are supposedly just "passed" from the control string to the handler _as they arrive_... @zadjii-msft I'll retract what I said earlier, if that's how it is. In that case, OSC 110 and 111 are **not** what I want. I obviously need OSC 10 and 11 instead. 😬 I guess that really, one way or another, I need to be able to force the background to a known value. I have been assuming that the windows console wouldn't suddenly learn how to have a different default bg/fg and would be limited to one of the 16 palette values ... I'm used to that by now. But I was thinking the defaults would be defaults in the global sense, or at least in the sense that's used in the window's menu... ![image](https://user-images.githubusercontent.com/192942/49671688-aea49600-fa36-11e8-8874-b51ab50ca2fb.png) So yeah. Anyway. From my perspective, I don't really _care_ if the background is black, blue, magenta -- but I do need **know for certain** which color index the background is. The goal is to avoid that awkward situation where I write text in the background color, when I thought I was writing highly visible pink or white on black ... And yes, the number one offender is the PowerShell blue background -- but people can (and do) change their defaults. I have had several co-workers with vision impairments and they all tend to force their consoles to black on a _white_ background.
claunia commented 2026-01-30 21:53:04 +00:00
Author
Owner
Copy Link

@oising commented on GitHub (Dec 10, 2018):

This probably belongs here as a reference for others finding this thread:

https://blogs.msdn.microsoft.com/commandline/2018/12/10/new-experimental-console-features/

@oising commented on GitHub (Dec 10, 2018): This probably belongs here as a reference for others finding this thread: https://blogs.msdn.microsoft.com/commandline/2018/12/10/new-experimental-console-features/ ![](https://msdnshared.blob.core.windows.net/media/2018/12/Experimental-settings.png)
claunia commented 2026-01-30 21:53:05 +00:00
Author
Owner
Copy Link

@Jaykul commented on GitHub (Dec 11, 2018):

Very exciting -- and makes OSC 10 and 11 even more critical 🤔 is there even a way to set those without the dialog right now? 🤨 And in particular querying.

How could an application update it's own knowledge of BackgroundColor or ForegroundColor to what the user has set in the dialog? I mean, PowerShell exposes those on the $Host.UI.RawUI -- but how could it read them? Is there any way it could know they'd been changed?

@Jaykul commented on GitHub (Dec 11, 2018): Very exciting -- and makes OSC 10 and 11 even more critical 🤔 is there even a way to set those without the dialog right now? 🤨 And in particular **querying**. How could an application update it's own knowledge of `BackgroundColor` or `ForegroundColor` to what the user has set in the dialog? I mean, PowerShell exposes those on the `$Host.UI.RawUI` -- but how could it read them? Is there any way it could know they'd been changed?
claunia commented 2026-01-30 21:53:05 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Dec 11, 2018):

Those aren't modifiable or queryable at the moment using VT. It's a pretty early experimental feature at the moment, so I'll make sure to bump up the priority of changing and querying those values, but that will all probably fall under the scope of the aforementioned deliverable.

@zadjii-msft commented on GitHub (Dec 11, 2018): Those aren't modifiable or queryable at the moment using VT. It's a pretty early experimental feature at the moment, so I'll make sure to bump up the priority of changing and querying those values, but that will all probably fall under the scope of the aforementioned deliverable.
claunia commented 2026-01-30 21:53:05 +00:00
Author
Owner
Copy Link

@oising commented on GitHub (Mar 2, 2019):

Any progress here, @zadjii-msft ?

@oising commented on GitHub (Mar 2, 2019): Any progress here, @zadjii-msft ?
claunia commented 2026-01-30 21:53:05 +00:00
Author
Owner
Copy Link

@Jaykul commented on GitHub (May 14, 2019):

OK, since we have source ... I'm going to try doing this myself, but I need a little help.

Over on Jaykul/SettableDefaultColors I've got a working version of OSC 10 and 11 for conhost, but although they clearly have some effect, I feel like they're not working properly in Terminal, and I'm not sure what I missed.

Can anyone either:

  • Give me a pointer on how to get C++ debugging to work when I launch the Terminal app (it works fine if I launch host.exe, but not Cascadia).
  • Take a look at the list of changes in 0f934b and tell me where I missed implementing it (there are a lot of places to implement this).
@Jaykul commented on GitHub (May 14, 2019): OK, since we have source ... I'm going to try doing this myself, but I need a little help. Over on [Jaykul/SettableDefaultColors](https://github.com/Jaykul/Terminal/tree/dev/jaykul/SettableDefaultColors) I've got a working version of OSC 10 and 11 for conhost, but although they _clearly_ have some effect, I feel like they're not working properly in Terminal, and I'm not sure what I missed. Can anyone either: - Give me a pointer on how to get C++ debugging to work when I launch the Terminal app (it works fine if I launch host.exe, but not Cascadia). - Take a look at the [list of changes in 0f934b](https://github.com/Jaykul/Terminal/commit/0f934b0cd8912bc9e495d2031b894e64648a701c) and tell me where I missed implementing it (there are **a lot** of places to implement this).
claunia commented 2026-01-30 21:53:06 +00:00
Author
Owner
Copy Link

@mKay00 commented on GitHub (May 14, 2019):

You have to set to Native only in the properties for debugging: https://github.com/microsoft/Terminal/blob/master/doc/Debugging.md

@mKay00 commented on GitHub (May 14, 2019): You have to set to Native only in the properties for debugging: https://github.com/microsoft/Terminal/blob/master/doc/Debugging.md
claunia commented 2026-01-30 21:53:06 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (May 14, 2019):

@Jaykul great work with that - that's pretty much everything that I'd think you'd need.

If it doesn't work for the terminal, I'd presume that for some reason, the color change isn't making it's way through conpty.

I'd think that the code you correctly grabbed to return false in conpty mode (near this line) should make sure the sequence is re-emitted through conpty. I'd set a breakpoint in TerminalCore/TerminalDispatch::SetDefaultForeground to make sure that the SetDefaultForeground handler is actually being called. It's possible that returning false doesn't actually bubble VT sequences up as I think they do

@zadjii-msft commented on GitHub (May 14, 2019): @Jaykul great work with that - that's pretty much everything that I'd think you'd need. If it doesn't work for the terminal, I'd presume that for some reason, the color change isn't making it's way through conpty. I'd think that the code you correctly grabbed to return false in conpty mode (near [this line](https://github.com/Jaykul/Terminal/commit/0f934b0cd8912bc9e495d2031b894e64648a701c#diff-cf3594b5816cdce2d4d1dca2f8a1d462R1862)) _should_ make sure the sequence is re-emitted through conpty. I'd set a breakpoint in `TerminalCore/TerminalDispatch::SetDefaultForeground` to make sure that the `SetDefaultForeground` handler is actually being called. It's possible that returning false doesn't actually bubble VT sequences up as I think they do
claunia commented 2026-01-30 21:53:06 +00:00
Author
Owner
Copy Link

@Jaykul commented on GitHub (May 14, 2019):

@zadjii-msft in a related note ... in OutputStateMachineEngine::ActionOscDispatch (and the other Action*Dispatch handlers) there's this pattern of a double switch: first pass to parse, second pass to actually dispatch ...

Is there a reason it needs to be like that?

It prevents supporting arrays like xterm does: multiple sets of semicolon-delimited index;colorspec pairs after a single OSC 4 (basically, we need to dispatch multiple times, as we're parsing).

It also complicates querying (when the value is ? instead of an rgb value). Actually -- I guess we're going to need a whole other set of API calls for queries so which call to make will depend on the parse results! Actually, I'm not sure how to inject output from inside the state machine, is that even possible?

@Jaykul commented on GitHub (May 14, 2019): @zadjii-msft in a related note ... in `OutputStateMachineEngine::ActionOscDispatch` (and the other Action*Dispatch handlers) there's this pattern of a double switch: first pass to parse, second pass to actually dispatch ... Is there a reason it **needs** to be like that? It prevents supporting arrays like xterm does: multiple sets of semicolon-delimited `index;colorspec` pairs after a single OSC 4 (basically, we need to dispatch multiple times, as we're parsing). It also complicates querying (when the value is `?` instead of an rgb value). Actually -- I guess we're going to need a whole other set of API calls for queries so which call to make will depend on the parse results! Actually, I'm not sure how to inject output from inside the state machine, is that even possible?
claunia commented 2026-01-30 21:53:06 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (May 15, 2019):

There's definitely no reason that the parser needs to behave like that. You're right that it probably shouldn't, I just haven't gotten the time to go back and update that yet :P

There definitely is a way to respond back to the caller from the adapter. I can't remember the name of the call off the top of my head, something like Device Status Request or something, where we reply with the cursor position.

@zadjii-msft commented on GitHub (May 15, 2019): There's definitely no reason that the parser _needs_ to behave like that. You're right that it probably shouldn't, I just haven't gotten the time to go back and update that yet :P There definitely is a way to respond back to the caller from the adapter. I can't remember the name of the call off the top of my head, something like Device Status Request or something, where we reply with the cursor position.
claunia commented 2026-01-30 21:53:07 +00:00
Author
Owner
Copy Link

@Jaykul commented on GitHub (May 18, 2019):

I figured out why the terminal wasn't working. In order to change the background color of the Windows Terminal, you have to reset all the Acrylic stuff. I've got it working!

@Jaykul commented on GitHub (May 18, 2019): I figured out why the terminal wasn't working. In order to change the background color of the Windows Terminal, you have to reset all the Acrylic stuff. I've got it working!
claunia commented 2026-01-30 21:53:07 +00:00
Author
Owner
Copy Link

@DHowett-MSFT commented on GitHub (Jul 3, 2019):

Hey hey, this is now available in the inbox windows console as of insider build 18932! Congrats!

@DHowett-MSFT commented on GitHub (Jul 3, 2019): Hey hey, this is now available in the inbox windows console as of insider build 18932! Congrats!
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 Issue-Feature Product-Conhost Resolution-Fix-Available
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#467
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?