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

v2 console: unblocked WriteFile sets lpNumberOfBytesWritten incorrectly #36

New Issue
Closed
opened 2026-01-30 21:40:14 +00:00 by claunia · 3 comments
claunia commented 2026-01-30 21:40:14 +00:00
Owner
Copy Link

Originally created by @rprichard on GitHub (Dec 17, 2017).

A console can be frozen/paused if there is text selection (or IIRC if the user hits the Pause key in certain input modes). In that case, WriteConsole/WriteFile calls are blocked until the console is unfrozen.

With ASCII output, when a WriteFile call successfully completes that had initially been blocked, WriteFile reports the number of bytes written (via lpNumberOfBytesWritten) as nNumberOfBytesToWrite * 2 rather than nNumberOfBytesToWrite. I'd guess that it's reporting the size, in bytes, of the output after a conversion to UTF-16. I expect it to use an unconverted amount.

WriteConsoleA works as expected, as does WriteFile in either a legacy console or with the v2 console in Windows 10.0.15063.

winpty uses the "Select All" command to synchronize with a stream of output into a console, so programs running winpty hit this WriteFile change frequently. In particular, the print statement in Python 2.7 can fail, which is causing problems for people:

upstream winpty bug: https://github.com/rprichard/winpty/issues/134

Affected Windows versions:

  • 10.0.16257.1000
  • 10.0.16299.19
  • 10.0.16299.125

Repro 1 (Python 2.7):

test.py :

import sys
for i in range(20000):
    sys.stdout.write(b'.')
    sys.stdout.flush()

Run the script, then press Ctrl-A to select all text before it finishes. Then press ESC. The output looks like this:

C:\Users\rprichard>c:\Python27\python.exe test.py
.................................................................................
.................................................................................
.................................................................................
......................Traceback (most recent call last):
  File "test.py", line 3, in <module>
    sys.stdout.write(b'.')
IOError: [Errno 0] Error

Repro 2 (C++11)

#pragma comment(lib, "user32.lib") 

#include <windows.h>
#include <stdio.h>
#include <string.h>

#include <thread>

// Send "Select All", then spawn a thread to hit ESC a moment later.
static void start_selection() {
    const HWND hwnd = GetConsoleWindow();
    const int SC_CONSOLE_SELECT_ALL = 0xFFF5;
    SendMessage(hwnd, WM_SYSCOMMAND, SC_CONSOLE_SELECT_ALL, 0);
    auto press_escape = std::thread([=]() {
        Sleep(300);
        SendMessage(hwnd, WM_CHAR, 27, 0x00010001);
    });
    press_escape.detach();
}

template <typename T>
static void do_write_test(
        const char *api_name,
        T *api_ptr,
        bool use_selection) {
    if (use_selection) {
        start_selection();
    }    
    char buf[] = "1234\n";
    DWORD actual = 0;
    const BOOL ret = api_ptr(
        GetStdHandle(STD_OUTPUT_HANDLE),
        buf, strlen(buf), &actual, NULL);
    const DWORD last_error = GetLastError();
    printf("%s: %s returned %d: actual=%u LastError=%u (%s)\n",
        ((ret && actual == strlen(buf)) ? "SUCCESS" : "ERROR"),
        api_name, ret, actual, last_error,
        use_selection ? "select" : "no-select");
}

int main() {
    // Passes: Unblocked WriteConsoleA.
    do_write_test("WriteConsoleA", WriteConsoleA, false);

    // Passes: Blocked WriteConsoleA.
    do_write_test("WriteConsoleA", WriteConsoleA, true);

    // Passes: Unblocked WriteFile.
    do_write_test("WriteFile", WriteFile, false);

    // Fails: Blocked WriteFile. WriteFile returns the number of bytes
    // multiplied by 2 (the size of the UTF-16 output in bytes?).
    do_write_test("WriteFile", WriteFile, true);

    return 0;
}

Bad output:

C:\Users\rprichard>cl /nologo consolebug.cpp
consolebug.cpp

C:\Users\rprichard>consolebug
1234
SUCCESS: WriteConsoleA returned 1: actual=5 LastError=0 (no-select)
1234
SUCCESS: WriteConsoleA returned 1: actual=5 LastError=0 (select)
1234
SUCCESS: WriteFile returned 1: actual=5 LastError=0 (no-select)
1234
ERROR: WriteFile returned 1: actual=10 LastError=0 (select)

Good output:

C:\Users\rprichard>\\nas\nas\SambaShared\consolebug.exe
1234
SUCCESS: WriteConsoleA returned 1: actual=5 LastError=0 (no-select)
1234
SUCCESS: WriteConsoleA returned 1: actual=5 LastError=0 (select)
1234
SUCCESS: WriteFile returned 1: actual=5 LastError=0 (no-select)
1234
SUCCESS: WriteFile returned 1: actual=5 LastError=0 (select)

Edit: update test to reflect that (a) WriteFile does the right thing when it isn't blocked by selection, and (b) WriteConsoleA works in either case.

Originally created by @rprichard on GitHub (Dec 17, 2017). A console can be frozen/paused if there is text selection (or IIRC if the user hits the Pause key in certain input modes). In that case, `WriteConsole`/`WriteFile` calls are blocked until the console is unfrozen. With ASCII output, when a `WriteFile` call successfully completes that had initially been blocked, `WriteFile` reports the number of bytes written (via `lpNumberOfBytesWritten`) as `nNumberOfBytesToWrite * 2` rather than `nNumberOfBytesToWrite`. I'd guess that it's reporting the size, in bytes, of the output after a conversion to UTF-16. I expect it to use an unconverted amount. `WriteConsoleA` works as expected, as does `WriteFile` in either a legacy console or with the v2 console in Windows 10.0.15063. winpty uses the "Select All" command to synchronize with a stream of output into a console, so programs running winpty hit this `WriteFile` change frequently. In particular, the `print` statement in Python 2.7 can fail, which is causing problems for people: - https://github.com/platformio/platformio-atom-ide-terminal/issues/422 - https://github.com/Microsoft/vscode/issues/40199 - https://github.com/platformio/platformio-vscode-ide/issues/61 upstream winpty bug: https://github.com/rprichard/winpty/issues/134 Affected Windows versions: - 10.0.16257.1000 - 10.0.16299.19 - 10.0.16299.125 # Repro 1 (Python 2.7): test.py : ```python import sys for i in range(20000): sys.stdout.write(b'.') sys.stdout.flush() ``` Run the script, then press Ctrl-A to select all text before it finishes. Then press ESC. The output looks like this: ``` C:\Users\rprichard>c:\Python27\python.exe test.py ................................................................................. ................................................................................. ................................................................................. ......................Traceback (most recent call last): File "test.py", line 3, in <module> sys.stdout.write(b'.') IOError: [Errno 0] Error ``` # Repro 2 (C++11) ```c++ #pragma comment(lib, "user32.lib") #include <windows.h> #include <stdio.h> #include <string.h> #include <thread> // Send "Select All", then spawn a thread to hit ESC a moment later. static void start_selection() { const HWND hwnd = GetConsoleWindow(); const int SC_CONSOLE_SELECT_ALL = 0xFFF5; SendMessage(hwnd, WM_SYSCOMMAND, SC_CONSOLE_SELECT_ALL, 0); auto press_escape = std::thread([=]() { Sleep(300); SendMessage(hwnd, WM_CHAR, 27, 0x00010001); }); press_escape.detach(); } template <typename T> static void do_write_test( const char *api_name, T *api_ptr, bool use_selection) { if (use_selection) { start_selection(); } char buf[] = "1234\n"; DWORD actual = 0; const BOOL ret = api_ptr( GetStdHandle(STD_OUTPUT_HANDLE), buf, strlen(buf), &actual, NULL); const DWORD last_error = GetLastError(); printf("%s: %s returned %d: actual=%u LastError=%u (%s)\n", ((ret && actual == strlen(buf)) ? "SUCCESS" : "ERROR"), api_name, ret, actual, last_error, use_selection ? "select" : "no-select"); } int main() { // Passes: Unblocked WriteConsoleA. do_write_test("WriteConsoleA", WriteConsoleA, false); // Passes: Blocked WriteConsoleA. do_write_test("WriteConsoleA", WriteConsoleA, true); // Passes: Unblocked WriteFile. do_write_test("WriteFile", WriteFile, false); // Fails: Blocked WriteFile. WriteFile returns the number of bytes // multiplied by 2 (the size of the UTF-16 output in bytes?). do_write_test("WriteFile", WriteFile, true); return 0; } ``` Bad output: ``` C:\Users\rprichard>cl /nologo consolebug.cpp consolebug.cpp C:\Users\rprichard>consolebug 1234 SUCCESS: WriteConsoleA returned 1: actual=5 LastError=0 (no-select) 1234 SUCCESS: WriteConsoleA returned 1: actual=5 LastError=0 (select) 1234 SUCCESS: WriteFile returned 1: actual=5 LastError=0 (no-select) 1234 ERROR: WriteFile returned 1: actual=10 LastError=0 (select) ``` Good output: ``` C:\Users\rprichard>\\nas\nas\SambaShared\consolebug.exe 1234 SUCCESS: WriteConsoleA returned 1: actual=5 LastError=0 (no-select) 1234 SUCCESS: WriteConsoleA returned 1: actual=5 LastError=0 (select) 1234 SUCCESS: WriteFile returned 1: actual=5 LastError=0 (no-select) 1234 SUCCESS: WriteFile returned 1: actual=5 LastError=0 (select) ``` Edit: update test to reflect that (a) `WriteFile` does the right thing when it isn't blocked by selection, and (b) `WriteConsoleA` works in either case.
claunia added the Product-Conhost label 2026-01-30 21:40:14 +00:00
claunia commented 2026-01-30 21:40:15 +00:00
Author
Owner
Copy Link

@rprichard commented on GitHub (Dec 17, 2017):

@miniksa @zadjii-msft I assume this is something you'll want to look at?

@rprichard commented on GitHub (Dec 17, 2017): @miniksa @zadjii-msft I assume this is something you'll want to look at?
claunia commented 2026-01-30 21:40:15 +00:00
Author
Owner
Copy Link

@khouzam commented on GitHub (Dec 17, 2017):

Thanks @rprichard,

We've addressed the issue recently it's https://github.com/Microsoft/vscode/issues/40199. We weren't doing the proper character conversion back to A when the call was not processed immediately.

@khouzam commented on GitHub (Dec 17, 2017): Thanks @rprichard, We've addressed the issue recently it's https://github.com/Microsoft/vscode/issues/40199. We weren't doing the proper character conversion back to A when the call was not processed immediately.
claunia commented 2026-01-30 21:40:16 +00:00
Author
Owner
Copy Link

@DHowett-MSFT commented on GitHub (Nov 21, 2018):

It looks like this fix shipped. Thanks for reporting it!

@DHowett-MSFT commented on GitHub (Nov 21, 2018): It looks like this fix shipped. Thanks for reporting it!
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 Product-Conhost
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#36
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?