Re: [MiNT] Keyboard-problem with XaAES helmut-branch

["Helmut Karlowski" <helmut.karlowski@ish.de>]

Am 27.06.2010, 18:39 Uhr, schrieb Jo Even Skarstein <joska@online.no>:


> > This is STOP. Try to send SIGCONT with the taskmanager by clicking on
> > the wake-button.
>
> Sorry, it's 's' not 'S'. Tried SIGCONT but it did not make a difference.

Then it's select, which is ok. Next version will have longer entries for  
STATE.

> > It could be stopped when it does IO and is not allowed to. This is
> > possible when pgrp does not fit console-owner. And here I've indeed
> > changed something.
>
> Can you please explain what pgrp is? This is not something I've had to

This is related to job-control. When a proc uses tty it has to own it  
(open()). All processes with same pgrp may use this tty as well. When you  
start a process from a shell asynchronously (with &), and it does tty-IO  
it may get stopped by the system (signals TTIN/TTOU). Once upon a time I  
knew more about this ...

But likely this does not apply to taskbar as it is in select-mode.

> relate to in MiNT before. According to the Task Manager, all processes
> have pgrp == pid, except the kernel threads which all have pgrp == 0.

On my TT currently toswin has pid 13 and pgrp 1 (the one of XaAES). On  
aranym taskbar also has the pgrp of XaAES as all XaAES-clients have.

> > The kernel does this. It switches modes in some read-function depending
>
> Can I check for this in Taskbar some way?

You need to access the tty-struct in the kernel. In XaAES it is:

  tty = (struct tty *)C.Aes->p->p_fd->ofiles[C.KBD_dev]->devinfo;

I don't know if you can access the proc-structure of taskbar from inside  
taskbar - maybe by using the /proc-interface as ps does.

> > When a keyboard-event is waited for by select, it gets the current mode
> > on return, not necessarily the one that fits the caller.
>
> Taskbar does indeed wait for keyboard events. Disabling that does not

Using Bconin?

> make a difference.

Hm ...


> > Maybe you changed some bits?
>
> I most definitely have ;-) I will send you the current version later
> tonight.

Ok - I wait for the third announcement ;)

But I would first like to know why your pgrps are different.

--
Helmut Karlowski