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

ReadConsoleW not handling tab completion correctly #1409

New Issue
Closed
opened 2026-01-30 22:25:22 +00:00 by claunia · 6 comments
claunia commented 2026-01-30 22:25:22 +00:00
Owner
Copy Link

Originally created by @RadAd on GitHub (May 30, 2019).

Environment

Microsoft Windows [Version 10.0.17763.503]

Steps to reproduce

#include <Windows.H>

int main()
{
    HANDLE hStdInput = GetStdHandle(STD_INPUT_HANDLE);

    WCHAR buf[1024];
    DWORD read = 0;
    CONSOLE_READCONSOLE_CONTROL crc = {};
    crc.nLength = sizeof(crc);
    WCHAR c = L'\t';
    crc.dwCtrlWakeupMask = 1 << c;
    ReadConsoleW(hStdInput, buf, ARRAYSIZE(buf), &read, &crc);

    HANDLE hStdOutput = GetStdHandle(STD_OUTPUT_HANDLE);

    DWORD written = 0;
    WriteConsoleW(hStdOutput, L"\n", 1, &written, NULL);
    WriteConsoleW(hStdOutput, buf, read, &written, NULL);

    return 0;
}

Run the above program and type in (no return): abcdef
Move the cursor back 3 places.
Press the TAB key

Expected behavior

I expected buf to contain abc\tdef

Actual behavior

Actually buf to contain abc\tef

Notice that the tab character is in the correct place but it seems to have overwritten the character rather than inserted it. The length, however, has been increased by one but the last character is a space which wasn't typed in. This makes it impossible to correctly handling tab completion.

Originally created by @RadAd on GitHub (May 30, 2019). # Environment Microsoft Windows [Version 10.0.17763.503] # Steps to reproduce ``` #include <Windows.H> int main() { HANDLE hStdInput = GetStdHandle(STD_INPUT_HANDLE); WCHAR buf[1024]; DWORD read = 0; CONSOLE_READCONSOLE_CONTROL crc = {}; crc.nLength = sizeof(crc); WCHAR c = L'\t'; crc.dwCtrlWakeupMask = 1 << c; ReadConsoleW(hStdInput, buf, ARRAYSIZE(buf), &read, &crc); HANDLE hStdOutput = GetStdHandle(STD_OUTPUT_HANDLE); DWORD written = 0; WriteConsoleW(hStdOutput, L"\n", 1, &written, NULL); WriteConsoleW(hStdOutput, buf, read, &written, NULL); return 0; } ``` Run the above program and type in (no return): ```abcdef``` Move the cursor back 3 places. Press the ```TAB``` key # Expected behavior I expected ```buf``` to contain ```abc\tdef``` # Actual behavior Actually ```buf``` to contain ```abc\tef ``` Notice that the tab character is in the correct place but it seems to have overwritten the character rather than inserted it. The length, however, has been increased by one but the last character is a space which wasn't typed in. This makes it impossible to correctly handling tab completion.
claunia added the Product-ConhostResolution-By-DesignIssue-BugArea-Server labels 2026-01-30 22:25:22 +00:00
claunia commented 2026-01-30 22:25:23 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (May 30, 2019):

Is this a regression from a previous version of Windows?

I'm pretty sure this is by design with the Console API. Sounds to me like you want to set ENABLE_INSERT_MODE with SetConsoleMode.

@zadjii-msft commented on GitHub (May 30, 2019): Is this a regression from a previous version of Windows? I'm pretty sure this is by design with the Console API. Sounds to me like you want to set `ENABLE_INSERT_MODE` with [SetConsoleMode](https://docs.microsoft.com/en-us/windows/console/setconsolemode).
claunia commented 2026-01-30 22:25:24 +00:00
Author
Owner
Copy Link

@DHowett-MSFT commented on GitHub (May 30, 2019):

Would you mind testing this after enabling [ ] Use legacy console?

@DHowett-MSFT commented on GitHub (May 30, 2019): Would you mind testing this after enabling `[ ] Use legacy console`?
claunia commented 2026-01-30 22:25:24 +00:00
Author
Owner
Copy Link

@RadAd commented on GitHub (May 31, 2019):

I not sure if it is a regression, just something I have noticed now. I think the CONSOLE_READCONSOLE_CONTROL was undocumented for a long time.

ENABLE_INSERT_MODE on or off doesn't change the behaviour.

Use legacy console has no effect either.

@RadAd commented on GitHub (May 31, 2019): I not sure if it is a regression, just something I have noticed now. I think the ```CONSOLE_READCONSOLE_CONTROL``` was undocumented for a long time. ```ENABLE_INSERT_MODE``` on or off doesn't change the behaviour. ```Use legacy console``` has no effect either.
claunia commented 2026-01-30 22:25:24 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Jun 6, 2019):

We're making this a bug to make sure that it hasn't regressed since previous versions.

I believe @dhowett-msft menitoned something about a rouge space character at the end of the buffer in this scenario - we should make sure it hasn't changed since previous versions

@zadjii-msft commented on GitHub (Jun 6, 2019): We're making this a bug to make sure that it hasn't regressed since previous versions. I believe @dhowett-msft menitoned something about a rouge space character at the end of the buffer in this scenario - we should make sure it hasn't changed since previous versions
claunia commented 2026-01-30 22:25:25 +00:00
Author
Owner
Copy Link

@DHowett-MSFT commented on GitHub (Aug 28, 2019):

Alright, I've gone and tested a wide gamut of console hosts.

version tab overwrites extra space at end of buffer
19h1 ✔️ ✔️
rs5 ✔️ ✔️
rs4 ✔️ ✔️
legacy v1 ✔️ ✔️

As unfortunate as this might be, it looks like it's not a regression. Unfortunately, with the Win32 Console APIs we don't really have the flexibility to change the little details like this. There's a very long tail of applications that could be depending on the buffer size manipulations we do here.

This code dates back as far as I can find: Windows 2000. It doesn't exist in NT 3.51 because CtrlWakeupMask doesn't exist in NT 3.51.

    if (CookedReadData->CtrlWakeupMask != 0 &&
        Char < L' ' && (CookedReadData->CtrlWakeupMask & (1 << Char))) {
        *CookedReadData->BufPtr = Char;
        CookedReadData->BytesRead += sizeof(WCHAR);
        CookedReadData->BufPtr += 1;
        CookedReadData->CurrentPosition += 1;
        CookedReadData->ControlKeyState = KeyState;
        return TRUE;
    }

Today, it looks like this.

974e95ebf7/src/host/readDataCooked.cpp (L495-L503)

When we're processing a cooked read, and the incoming character happens to be in the CTRL wakeup mask, we'll always place it at the current position in the incoming buffer, claim we read an additional character, and advance our position.

I hate to close bad things as "By Design," but I think this one has too much stuff piled on top of it for us to ever change it now.

@DHowett-MSFT commented on GitHub (Aug 28, 2019): Alright, I've gone and tested a wide gamut of console hosts. |version|tab overwrites|extra space at end of buffer| |-|-|-| |19h1|✔️|✔️| |rs5|✔️|✔️| |rs4|✔️|✔️| |legacy v1|✔️|✔️| As unfortunate as this might be, it looks like it's _not a regression_. Unfortunately, with the Win32 Console APIs we don't really have the flexibility to change the little details like this. There's a _very_ long tail of applications that could be depending on the buffer size manipulations we do here. This code dates back as far as I can find: Windows 2000. It doesn't exist in NT 3.51 because CtrlWakeupMask doesn't exist in NT 3.51. ``` if (CookedReadData->CtrlWakeupMask != 0 && Char < L' ' && (CookedReadData->CtrlWakeupMask & (1 << Char))) { *CookedReadData->BufPtr = Char; CookedReadData->BytesRead += sizeof(WCHAR); CookedReadData->BufPtr += 1; CookedReadData->CurrentPosition += 1; CookedReadData->ControlKeyState = KeyState; return TRUE; } ``` Today, it looks like this. https://github.com/microsoft/terminal/blob/974e95ebf7704b08da51c1f728ab2bc06108881f/src/host/readDataCooked.cpp#L495-L503 When we're processing a cooked read, and the incoming character happens to be in the CTRL wakeup mask, we'll _always_ place it at the current position in the incoming buffer, claim we read an additional character, and advance our position. I hate to close bad things as "By Design," but I think this one has too much stuff piled on top of it for us to ever change it now.
claunia commented 2026-01-30 22:25:25 +00:00
Author
Owner
Copy Link

@RadAd commented on GitHub (Aug 29, 2019):

Im not surprised its not a regression.

My current workaround is to read the lost character from the console.

I assume this was never picked up because cmd.exe throws away everything after the tab anyway.

@RadAd commented on GitHub (Aug 29, 2019): Im not surprised its not a regression. My current workaround is to read the lost character from the console. I assume this was never picked up because cmd.exe throws away everything after the tab anyway.
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-Server Issue-Bug Product-Conhost Resolution-By-Design
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#1409
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?