starred/terminal
1
0
Fork 0
mirror of https://github.com/microsoft/terminal.git synced 2026-09-25 08:24:54 +00:00
Code Issues 19.0k Packages Projects Releases 20 Wiki Activity

"invalid bounds arguments passed to std::clamp" when resizing window to zero height #17259

New Issue
Closed
opened 2026-01-31 05:37:02 +00:00 by claunia · 7 comments
claunia commented 2026-01-31 05:37:02 +00:00
Owner
Copy Link

Originally created by @Predelnik on GitHub (Apr 16, 2022).

Windows Terminal version

1.12.10733.0

Windows build number

10.0.22000

Other Software

No response

Steps to reproduce

Launch OpenConsole.exe compiled in debug mode, drag bottom of the window to the very top.

Expected Behavior

No assertion

Actual Behavior

Assertion from from STL "invalid bounds arguments passed to std::clamp" on the line

clampMe.Top = std::clamp(clampMe.Top, Top(), BottomInclusive());

in Viewport Viewport::Clamp

Call stack 
 	ucrtbased.dll!00007ff9248a7fbc()	Unknown
 	ucrtbased.dll!00007ff9248a7ef4()	Unknown
 	ucrtbased.dll!00007ff9248a7d90()	Unknown
>	OpenConsole.exe!std::clamp>(const short & _Val, const short & _Min_val, const short & _Max_val, std::less _Pred) Line 10188	C++
 	OpenConsole.exe!std::clamp(const short & _Val, const short & _Min_val, const short & _Max_val) Line 10208	C++
 	OpenConsole.exe!Microsoft::Console::Types::Viewport::Clamp(const Microsoft::Console::Types::Viewport & other) Line 245	C++
 	OpenConsole.exe!Microsoft::Console::Render::Renderer::TriggerRedrawCursor(const _COORD * const pcoord) Line 284	C++
 	OpenConsole.exe!TextBuffer::TriggerRedrawCursor(const _COORD position) Line 992	C++
 	OpenConsole.exe!Cursor::_RedrawCursorAlways() Line 190	C++
 	OpenConsole.exe!Cursor::SetIsOn(const bool fIsOn) Line 103	C++
 	OpenConsole.exe!Microsoft::Console::CursorBlinker::TimerRoutine(SCREEN_INFORMATION & ScreenInfo) Line 153	C++
 	OpenConsole.exe!CursorTimerRoutineWrapper(_TP_CALLBACK_INSTANCE * __formal, void * __formal, _TP_TIMER * __formal) Line 24	C++
 	ntdll.dll!TppTimerpExecuteCallback()	Unknown
 	ntdll.dll!TppWorkerThread()	Unknown
 	kernel32.dll!00007ff9feee54e0()	Unknown
 	ntdll.dll!RtlUserThreadStart()	Unknown

Looks like it happens because SCREEN_INFORMATION::_InternalSetViewportSize increases viewport's Top in srNewViewport.Top -= DeltaY; (DeltaY is negative) line without any check to prevent Top becoming more than Bottom which sadly can happen. Clamp cannot work when Top is more than Bottom

Originally created by @Predelnik on GitHub (Apr 16, 2022). ### Windows Terminal version 1.12.10733.0 ### Windows build number 10.0.22000 ### Other Software _No response_ ### Steps to reproduce Launch OpenConsole.exe compiled in debug mode, drag bottom of the window to the very top. ### Expected Behavior No assertion ### Actual Behavior Assertion from from STL `"invalid bounds arguments passed to std::clamp"` on the line ```cpp clampMe.Top = std::clamp(clampMe.Top, Top(), BottomInclusive()); ``` in `Viewport Viewport::Clamp` <details> <summary>Call stack</summary> <pre> ucrtbased.dll!00007ff9248a7fbc() Unknown ucrtbased.dll!00007ff9248a7ef4() Unknown ucrtbased.dll!00007ff9248a7d90() Unknown > OpenConsole.exe!std::clamp<short,std::less<void>>(const short & _Val, const short & _Min_val, const short & _Max_val, std::less<void> _Pred) Line 10188 C++ OpenConsole.exe!std::clamp<short>(const short & _Val, const short & _Min_val, const short & _Max_val) Line 10208 C++ OpenConsole.exe!Microsoft::Console::Types::Viewport::Clamp(const Microsoft::Console::Types::Viewport & other) Line 245 C++ OpenConsole.exe!Microsoft::Console::Render::Renderer::TriggerRedrawCursor(const _COORD * const pcoord) Line 284 C++ OpenConsole.exe!TextBuffer::TriggerRedrawCursor(const _COORD position) Line 992 C++ OpenConsole.exe!Cursor::_RedrawCursorAlways() Line 190 C++ OpenConsole.exe!Cursor::SetIsOn(const bool fIsOn) Line 103 C++ OpenConsole.exe!Microsoft::Console::CursorBlinker::TimerRoutine(SCREEN_INFORMATION & ScreenInfo) Line 153 C++ OpenConsole.exe!CursorTimerRoutineWrapper(_TP_CALLBACK_INSTANCE * __formal, void * __formal, _TP_TIMER * __formal) Line 24 C++ ntdll.dll!TppTimerpExecuteCallback() Unknown ntdll.dll!TppWorkerThread() Unknown kernel32.dll!00007ff9feee54e0() Unknown ntdll.dll!RtlUserThreadStart() Unknown </pre> </details> Looks like it happens because `SCREEN_INFORMATION::_InternalSetViewportSize` increases viewport's `Top` in `srNewViewport.Top -= DeltaY;` (`DeltaY` is negative) line without any check to prevent `Top` becoming more than `Bottom` which sadly can happen. Clamp cannot work when `Top` is more than `Bottom`
claunia commented 2026-01-31 05:37:04 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Apr 18, 2022):

For my own reference, which commit are you building off of? I just wanna know how far back we'll have to service this fix 😅

@zadjii-msft commented on GitHub (Apr 18, 2022): For my own reference, which commit are you building off of? I just wanna know how far back we'll have to service this fix 😅
claunia commented 2026-01-31 05:37:04 +00:00
Author
Owner
Copy Link

@Predelnik commented on GitHub (Apr 18, 2022):

I'm building recent main (v1.10.1933.0-605-g41ef5554f) and haven't tried to figure out when exactly it appeared unfortunately.

@Predelnik commented on GitHub (Apr 18, 2022): I'm building recent `main` (`v1.10.1933.0-605-g41ef5554f`) and haven't tried to figure out when exactly it appeared unfortunately.
claunia commented 2026-01-31 05:37:05 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Apr 18, 2022):

uh, https://github.com/microsoft/terminal/releases/tag/v1.10.1933.0 is like, 10 months old... If this goes that far back, well, we're in some serious trouble

edit: just guesses

@zadjii-msft commented on GitHub (Apr 18, 2022): uh, https://github.com/microsoft/terminal/releases/tag/v1.10.1933.0 is like, 10 months old... If this goes that far back, well, we're in some serious trouble edit: just guesses * https://github.com/microsoft/terminal/blame/41ef5554f5cb946c4a48e2b8db791265c7217fce/src/renderer/base/renderer.cpp#L266 - #8664 added the Clamp call to `Renderer::TriggerRedrawCursor`. There wasn't one there before that.
claunia commented 2026-01-31 05:37:05 +00:00
Author
Owner
Copy Link

@j4james commented on GitHub (Apr 18, 2022):

The more I think about, the more I think that call to view.Clamp is fundamentally wrong anyway. The intention was to clip the cursor rect to the viewport, so you'd only be drawing the portion of the cursor that was visible, but that's not what Viewport::Clamp does. If the cursor is completely off screen, it's basically going to pull it back into view, forcing the renderer to redraw an area of the screen that doesn't actually need refreshing. That's probably not the end of the world, but it's not right.

For the same reason, I don't think this bounds error is the end of the world either. In a release build, the clamp is just going to fail to do anything, but that's no worse than the off screen clamping. It's just invalidating an area of the screen that doesn't need redrawing.

So while we definitely should be fixing this, I don't think you need to panic about backporting the fix, unless I'm missing something.

@j4james commented on GitHub (Apr 18, 2022): The more I think about, the more I think that call to `view.Clamp` is fundamentally wrong anyway. The intention was to clip the cursor rect to the viewport, so you'd only be drawing the portion of the cursor that was visible, but that's not what `Viewport::Clamp` does. If the cursor is completely off screen, it's basically going to pull it back into view, forcing the renderer to redraw an area of the screen that doesn't actually need refreshing. That's probably not the end of the world, but it's not right. For the same reason, I don't think this bounds error is the end of the world either. In a release build, the clamp is just going to fail to do anything, but that's no worse than the off screen clamping. It's just invalidating an area of the screen that doesn't need redrawing. So while we definitely should be fixing this, I don't think you need to panic about backporting the fix, unless I'm missing something.
claunia commented 2026-01-31 05:37:06 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Apr 29, 2022):

Oh oh oh maybe this is why #12739

Maybe we're clamping the cursor, drawing it there, and then not invalidating it at the clamped place. Idk. Just a thought while fixing this. EDIT: I was wrong. that one still repros.

@zadjii-msft commented on GitHub (Apr 29, 2022): Oh oh oh maybe this is why #12739 Maybe we're clamping the cursor, drawing it there, and then not invalidating it at the clamped place. Idk. Just a thought while fixing this. EDIT: I was wrong. that one still repros.
claunia commented 2026-01-31 05:37:06 +00:00
Author
Owner
Copy Link

@j4james commented on GitHub (Jan 11, 2023):

@zadjii-msft I think we may want to reopen this issue. I was just testing PR #14661, which kind of reverts your fix for this, and I wanted to make sure I hadn't reintroduced the bug. What I discovered was that the issue was never fully resolved - while we don't get a clamp assertion in the cursor code, we now just get it somewhere else instead! For me it occurred here:

4d27a05318/src/terminal/adapter/adaptDispatch.cpp (L219)

And I double checked with the commit from when PR #13001 was merged (79904361b3), and it was already failing then. It may not be obvious in some shells, but in a WSL bash shell it crashes straight away. In a cmd shell I had to type some command (e.g. CLS) while the window had a zero size in order to trigger the crash.

It's not a major problem, since it only happens in a debug build, but if we think it needs fixing we should probably reopen this, or at least open another issue if you prefer that.

As for how we fix it, I would have liked it if we could put a minimum size on the viewport, but I suspect that's not possible in conhost for legacy reasons. So failing that, a quick fix would be to replace the clamp calls with a min/max combo which won't assert. But my concern is that the VT architecture is kind of dependent on a non-zero screen size, so there's a fair chance this issue might just surface elsewhere at some point in the future. Maybe we just have to accept that.

@j4james commented on GitHub (Jan 11, 2023): @zadjii-msft I think we may want to reopen this issue. I was just testing PR #14661, which kind of reverts your fix for this, and I wanted to make sure I hadn't reintroduced the bug. What I discovered was that the issue was never fully resolved - while we don't get a clamp assertion in the cursor code, we now just get it somewhere else instead! For me it occurred here: https://github.com/microsoft/terminal/blob/4d27a05318f4e86b0918497ba6f980968725a1d2/src/terminal/adapter/adaptDispatch.cpp#L219 And I double checked with the commit from when PR #13001 was merged (79904361b345b68f4612d41b390c4eb3f3d18fae), and it was already failing then. It may not be obvious in some shells, but in a WSL bash shell it crashes straight away. In a cmd shell I had to type some command (e.g. `CLS`) while the window had a zero size in order to trigger the crash. It's not a major problem, since it only happens in a debug build, but if we think it needs fixing we should probably reopen this, or at least open another issue if you prefer that. As for how we fix it, I would have liked it if we could put a minimum size on the viewport, but I suspect that's not possible in conhost for legacy reasons. So failing that, a quick fix would be to replace the clamp calls with a min/max combo which won't assert. But my concern is that the VT architecture is kind of dependent on a non-zero screen size, so there's a fair chance this issue might just surface elsewhere at some point in the future. Maybe we just have to accept that.
claunia commented 2026-01-31 05:37:06 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Jan 11, 2023):

Thanks for the write up @j4james. I'm gonna just promote that to another thread to try and keep the history a bit cleaner.

@zadjii-msft commented on GitHub (Jan 11, 2023): Thanks for the write up @j4james. I'm gonna just promote that to another thread to try and keep the history a bit cleaner.
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-Output Issue-Bug Needs-Tag-Fix Priority-0 Product-Conhost Resolution-Fix-Committed Severity-Crash
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#17259
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?