KeyPress crash

One of my friends got this one using my
psi-sneakin-20050622172553-win32 build.  From my investigation (and
his comments) he pressed 'y' after deciding not to resend the previous
line of text in a chat.

Psi had been open for 12-18 hours, and tabbed chat was disabled.  He
hadn't been using any special MUC features, (he just happened to be
using the MUC build)

The crash occurs at the end of PsiApplication::notify and deeper down
in ChatEdit::keyPressEvent.  I'm guessing something got deleted and
then accessed.

This is via a crash dump from DrWatson, so I can always open it back
up and check any variables at any time in the future.

The horrid stack trace:
 	ntdll.dll!77f82e3f() 	
 	ntdll.dll!77f58450() 	
 	ntdll.dll!77f944a8() 	
 	ntdll.dll!77f944a8() 	
 	ntdll.dll!77f944a8() 	
 	ntdll.dll!77f57d70() 	
 	ntdll.dll!77f58a3a() 	
 	ntdll.dll!77f944a8() 	
 	ntdll.dll!77f944a8() 	
 	ntdll.dll!77f57d70() 	
 	ntdll.dll!77f944a8() 	
 	ntdll.dll!77f57d70() 	
 	ntdll.dll!77f58a3a() 	

What is the release schedule for Psi 0.10?

Hi everybody.

Somebody asked me todau on my website about the 0.10 release of Psi. 
Unfortunately, I am not really aware of the latest release plan.

I've been following discussion and the only recent relevant piece of 
information on the release of Psi was from the delta maillist, where Justin 
mentioned QCA 2.0 beta1 (I think) will be shipped with 0.10. QCA is 2.0 
pre-beta1 (alpha?) right now.

On Monday 01 August 2005 14:19, Justin Karneges wrote:
> No further API changes are planned.  My intention is to use beta1 for the
> 0.10 development of Psi.

Well, I'm a bit confused since I had the feeling QCA2 is Qt4 only, and Psi 
0.10 will be Qt3.

I'm asking this here because:
1. Kevin is not online right now and I might forget about this later
2. Other people might be interested

Mircea
/IceRAM

--

-- 
Psi Forums Moderator/Bug Tracker Manager
Psi Windows Installer Maintainer/ArchLinux Package Maintainer
http://mircea.bardac.net

Re: What is the release schedule for Psi 0.10?

Mircea Bardac wrote:
> Somebody asked me todau on my website about the 0.10 release of Psi. 
> Unfortunately, I am not really aware of the latest release plan.
The current release plan is: finish fixing the bugs and release.

> I've been following discussion and the only recent relevant piece of 
> information on the release of Psi was from the delta maillist, where Justin 
> mentioned QCA 2.0 beta1 (I think) will be shipped with 0.10. QCA is 2.0 
> pre-beta1 (alpha?) right now.
Justin made a typo, QCA2 will be shipped with 0.11. :)

/K
--

-- 
Kevin Smith
Psi(.affinix.com) Jabber client project leader
PhD Research Student, Dept. Computer Science, University Of Exeter

Re: KeyPress crash

glad someone else is seeing crashes.  I've never been able to pin
anything down, but I didn't think it was unique to my system.

On 8/2/05, Norman Rasmussen <normanr@...> wrote:
> One of my friends got this one using my
> psi-sneakin-20050622172553-win32 build.  From my investigation (and
> his comments) he pressed 'y' after deciding not to resend the previous
> line of text in a chat.
> 
> Psi had been open for 12-18 hours, and tabbed chat was disabled.  He
> hadn't been using any special MUC features, (he just happened to be
> using the MUC build)
> 
> The crash occurs at the end of PsiApplication::notify and deeper down
> in ChatEdit::keyPressEvent.  I'm guessing something got deleted and
> then accessed.
> 
> This is via a crash dump from DrWatson, so I can always open it back
> up and check any variables at any time in the future.
> 
> The horrid stack trace:
>         ntdll.dll!77f82e3f()
>         ntdll.dll!77f58450()
>         ntdll.dll!77f944a8()
>         ntdll.dll!77f944a8()
>         ntdll.dll!77f944a8()
>         ntdll.dll!77f57d70()
>         ntdll.dll!77f58a3a()
>         ntdll.dll!77f944a8()
>         ntdll.dll!77f944a8()

Re: KeyPress crash

On Tuesday 02 August 2005 13:42, Norman Rasmussen wrote:
> One of my friends got this one using my
> psi-sneakin-20050622172553-win32 build.  From my investigation (and
> his comments) he pressed 'y' after deciding not to resend the previous
> line of text in a chat.
>
> The crash occurs at the end of PsiApplication::notify and deeper down
> in ChatEdit::keyPressEvent.  I'm guessing something got deleted and
> then accessed.

I have had rare crashes under Linux with the development version (Jun 10, 
2005). Unfortunately, it happened quite rare and I couldn't track them. Psi 
simply dissapeared from everywhere, leaving nothing behind.

This did happen while I was chatting with somebody. It never vanished from my 
face. It always happened when Psi received a message with the windows being 
not focused. I always felt I was loosing messages. I think I've seen the 
composing event in the taskbar, then everything vanished.

I find it pretty hard to link the crash to the content of the message, since 
it has no logic. I think the last message (the one I thought lost) got in the 
history - this happened a long time ago and I don't remember this well.

I've remembered all this because this weird crash was. It might be connected 
to what I've experienced... but.. don't really know.

Mircea
/IceRAM

--

-- 

Commenting your code

As much as I hate to admit to reading slashdot, the following article 
was linked earlier today. Everyone who codes should read it in my not 
very humble opinion, and I should probably come up with definite rules 
on some of the things discussed within.

Please make appropriate comments.

http://particletree.com/features/successful-strategies-for-commenting-your-code

/K

--

-- 
Kevin Smith
Psi(.affinix.com) Jabber client project leader
PhD Research Student, Dept. Computer Science, University Of Exeter

Re: Commenting your code

> and I should probably come up with definite rules
> on some of the things discussed within.

Actually keeping your code clean is more needed, in a project like
Psi, in which lots of people work with the same code. Not only
commenting your code, but organizing your functions (first public
functions, auxiliary functions later) and your classes help when some
other coders need to 'hack' your work, because they need it in some
other place. And good coments save a lot of time, that otherwise you
invert on finding out what makes each function...

If you're asking for opinions, I say we need those guidelines, and
start commenting not only our functions, but the procedures and
classes that are not and should be.

--

-- 
Francisco Joaquín Rodríguez Prados
Fachbereich Informatik,
Fachhochschule Darmstadt,
Darmstadt, Germany
E-mail: prados@...
JID: franqui@...

Re: KeyPress crash

On Tue, Aug 02, 2005 at 10:30:19PM +0300, Mircea Bardac wrote:
> On Tuesday 02 August 2005 13:42, Norman Rasmussen wrote:
> > One of my friends got this one using my
> > psi-sneakin-20050622172553-win32 build.  From my investigation (and
> > his comments) he pressed 'y' after deciding not to resend the previous
> > line of text in a chat.
> >
> > The crash occurs at the end of PsiApplication::notify and deeper down
> > in ChatEdit::keyPressEvent.  I'm guessing something got deleted and
> > then accessed.
> 
> I have had rare crashes under Linux with the development version (Jun 10, 
> 2005). Unfortunately, it happened quite rare and I couldn't track them. Psi 
> simply dissapeared from everywhere, leaving nothing behind.
> 
> This did happen while I was chatting with somebody. It never vanished from my 
> face. It always happened when Psi received a message with the windows being 
> not focused. I always felt I was loosing messages. I think I've seen the 
> composing event in the taskbar, then everything vanished.
> 
> I find it pretty hard to link the crash to the content of the message, since 
> it has no logic. I think the last message (the one I thought lost) got in the 
> history - this happened a long time ago and I don't remember this well.
> 
> I've remembered all this because this weird crash was. It might be connected 
> to what I've experienced... but.. don't really know.
> 

I got a some crashes recently too. But i thought maybe one of my local
patches caused it.

Re: Commenting your code

> If you're asking for opinions, I say we need those guidelines, and
> start commenting not only our functions, but the procedures and
> classes that are not and should be.

We have been enforcing this in our patches for a while now, as well as
well organized code.

cheers,
Remko

[Patch] Custom rostericons in tooltips

As we all know, Psi displays tooltips when hovering over the status icon 
(either in roster, or in chat dialog, or in groupchat dialog). They 
contain rostericons, but they always come from the default 
rostericonset, even if custom rostericonset is set for given contact.

My small patch (BTW: the first one using Darcs) deals with that problem. 
So, have fun!

-- 
Michał Jazłowiecki (michalj)
Psi Forum Moderator
Psi Tester & Patcher

As we all know, Psi displays tooltips when hovering over the status icon 
(either in roster, or in chat dialog, or in groupchat dialog). They 
contain rostericons, but they always come from the default 
rostericonset, even if custom rostericonset is set for given contact.

My small patch (BTW: the first one using Darcs) deals with that problem. 
So, have fun!

--

-- 
Michał Jazłowiecki (michalj)
Psi Forum Moderator
Psi Tester & Patcher