Lost keyboard input after suspend on Windows 10 PowerShell

[derailed]

I appears tcell v2.2.1 on Windows10 using powershell broke all keyboard inputs upon suspending and resuming an application. This behavior worked correctly of v.2.2.0.

[gdamore]

This does not seem to be a problem running the mouse.go demo application on Windows 10.

Can you share a bit more about what you're doing in the suspend?

[gdamore]

(For mouse.go, just press ctrl-z to "suspend".)

[derailed]

Thank you for the prompt reply Garrett!
I am doing a fork and exec and essentially running a shell so very close to what you're doing in mouse.go. I've checked out mouse.go on windows and you're correct it does seem to keep inputs upon resume. Also v2.2.1 works correctly on osx.
On windows, this worked correctly (minus cursor) on v2.2.0. I've just updated to v2.2.1 and my app is no longer responding to keyboard input upon resume. I haven't debugged it as of yet to see what's happening....

[gdamore]

The only thing I did in v2.2.1 for Windows different was to fix the cursor.

[derailed]

FYI I've reverted commits 3d0a and d20e and my app is now working correctly after suspend/resume on windows...

[gdamore]

<!-- /* Font Definitions */ @font-face {font-family:"Cambria Math"; panose-1:2 4 5 3 5 4 6 3 2 4;} @font-face {font-family:Calibri; panose-1:2 15 5 2 2 2 4 3 2 4;} /* Style Definitions */ p.MsoNormal, li.MsoNormal, div.MsoNormal {margin:0in; font-size:11.0pt; font-family:"Calibri",sans-serif;} a:link, span.MsoHyperlink {mso-style-priority:99; color:blue; text-decoration:underline;} .MsoChpDefault {mso-style-type:export-only;} @page WordSection1 {size:8.5in 11.0in; margin:1.0in 1.0in 1.0in 1.0in;} div.WordSection1 {page:WordSection1;} -->3d0a has nothing to do with Windows, and shouldn’t even be part of your binary.   (Assuming you mean d003d0a) However e37d20e is interesting.  I need to look more closely at this. Sent from Mail for Windows 10 From: Fernand GalianaSent: Monday, April 19, 2021 10:36 AMTo: gdamore/tcellCc: gdamore; CommentSubject: Re: [gdamore/tcell] Lost keyboard input after suspend on Windows 10 PowerShell... (#449) FYI I've reverted commits 3d0a and d20e and my app is now working correctly after suspend/resume on windows...—You are receiving this because you commented.Reply to this email directly, view it on GitHub, or unsubscribe. 

[derailed]

@gdamore Any update on this issue? Fixed windows but broke linux. Any ways out of the predicament? Thanks Garrett!

[gdamore]

Please have a look at my "tty" branch. I think it works well on both Windows and Linux now. I haven't tested macOS as I don't have a macOS system handy at this precise instant.

[derailed]

Tx Garrett!
I could not find a branch tty on this repo?? but went to the commit instead...

Sadly I don't think it's fixed ;(

1. Key bindings toast after resume on windows 10 Powershell.
2. On OSX the screen only shows the first line on resume, try to resize to force a redraw bombs on tty_unix.go:80

Please reopen so we can track this down as my app is currently in no mans land ;(

[gdamore]

Please update to v2.3.1. If that still doesn't work, let me know.

[derailed]

Thanks Garrett! I did try out v2.3.1 seems to have resolved the resume issue on OSX. Thank you for that!
But still a dud on windows ie upon resume, lost all my keyboard inputs ;( This is using powershell on windows 10.

[gdamore]

Does the _demos/mouse.go program work for you? -- if you press ctrl-z it should work, and exit should work, with normal keys working. I'm using Windows 10 with powershell myself.

[gdamore]

Make sure you aren't calling Fini(). Only Suspend(). Fini() will probably do things you don't want.

[derailed]

Tx for the suggestions Garret!

I think the problem lies Here
When the screen is disengage stopQ channel is closed thus stopping the event loop since the select will fire on that case. I think the issue here is we should stop the event poll when the Fini is called. But for windows Fini just turns around and calls disengage thus there is no clean delineation when the screen is suspended vs when we're done with the screen.

Note: I could be totally off here as I am my phone as more mem than the crap windows box I am on ;( Less than ideal to debug in this env but changing the line above to read <-s.quit and having Fini close(s.quit) instead seems to fix particular issue.

Does this make sense?