assertion failed "nNewWidth>0 && nNewHeight>0" in RealBuffer.cpp:1894 ConEmu 161022

[0anton]

Hi,

I experience ConEmu crashes after resizing ConEmu window.

Assertion ConEmu 161022 [32] PID=7808, TID=14216
RealBuffer.cpp:1894
nNewWidth>0 && nNewHeight>0

Versions

ConEmu build: 161022 x64
OS version: Windows 10 x64
Used shell: cygwin

Problem description

ConEmu crashes after console resize (e.g. full-screen / part-screen toggle)

Steps to reproduce

1. Start Conemu.
2. Create couple of shells in the screen split mode.
3. Open some files.
4. Toggle full-screen for one of them.

Actual results

ConEmu crashes.

Expected results

ConEmu continue to work.

Additional files

[Traces]
https://www.dropbox.com/s/q77q57047exc4mx/Trap-161022-7808.zip?dl=0

[Settings]
https://www.dropbox.com/s/2jlceqaicbcoyhd/ConEmu.xml?dl=0

[screenshots]
https://www.dropbox.com/s/fhdeezyip86k2tz/ScreenShot031.jpg?dl=0

[logs]

(will follow with the next crash)

[Maximus5]

That usually means something bad happened with conhost.exe.
Can you reproduce? Than attach LogFiles please.

[0anton]

Sure Max,

I'll provide logs with the next crash.

Kind regards,
Anton

On Tue, Nov 1, 2016 at 3:27 PM, Maximus5 notifications@github.com wrote:

That usually means something bad happened with conhost.exe.
Can you reproduce? Than attach LogFiles please.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#934 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AQgg5C44Tjbrxehh6I7AD2yh-ZZx4pBUks5q50xhgaJpZM4KmJPI
.

[0anton]

Hi Max,

As requested, please find the crash log here:

ConEmu-srv-19028.zip

Kind regards,
Anton

[Maximus5]

2 all

I need more information to reproduce the issue.

1. What are you doing exactly?
2. Show screenshots with status bar visible of course. I need at least these columns: Terminal modes, Console visible rectangle, Console buffer size, Cursor information, Console server PID.
3. I need all LogFiles created by this instance of ConEmu. Not only one selected file without any information what were you doing.

The issue is hard to fix because it depens on some race conditions in Microsoft's conhost.exe.

[Shredder121]

At the time I encountered my issue, it was easily reproducible.
You would just generate output until you were at the end of the buffer (scroll handle at the bottom of the scroll bar), and then clear.
But now, with a fresh install (both Windows and conemu) it doesn't happen.

So something, somewhere must break it, but I think it's not directly conemu.

[0anton]

Hi Max, The problem disappeared from itself. May be Windows updated cohost.exe file.. I will watch and report if if I have more details... Kind regards, Anton Golubev

…

On Thu, Dec 8, 2016 at 11:52 AM, Ruben Dijkstra ***@***.***> wrote: At the time I encountered my issue, it was easily reproducible. You would just generate output until you were at the end of the buffer (scroll handle at the bottom of the scroll bar), and then clear. But now, with a fresh install (both Windows and conemu) it doesn't happen. So something, somewhere must break it, but I think it's not *directly* conemu. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#934 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AQgg5JhZilA2A5ikjX3LIIobuaC6-rTqks5rF-FRgaJpZM4KmJPI> .

[m0n]

ConEmu build: 180206 x64
OS version: Windows 10 LTSB x64
Used shell: Far Manager, version 3.0 (build 5151) x64

I can reproduce this assertion 10/10 on my PC:

1. ConEmu64 -basic -log
2. FAR.exe - without this step everything is OK
3. wmic.exe diskdrive list brief /format:list
4. Command executed properly
5. wmic.exe
6. Assertion happens

The dump & logs of this sequence are here.

I noticed that starting wmic.exe modifies "Screen Buffer Size" to 1500*300 in a normal cmd.exe (with "Use legacy console")

Wow - I just realised that if You set "Use legacy console" to "on" then wmic.exe starts with an extra delay but no assertion happens! This is an answer to why this bug not appears on Windows 7...

[0anton]

I think this issue can be closed since much time passed by.