KTechlab

Thanks for the backtraces :) (how you made ktechlab crash is also useful,
even if you can't reproduce the crash...)

I've hopefully fixed two of the bugs you sent me (and discovered a new one
while investigating...), but I'm still unsure about what's going on in
KTechLab_bt2.txt - some sleep should help here (same applies to
potentiometer glitch).

> After compiling it succesfully with the later revision I've been testing
> it a little.
>
> I've found some crashes (backtraces attached) and some graphic glitches,
> for example, if you put a potentiomenter, wire it to something, and rotate
> it, then one pin is misaligned with the wire.
>
> I will continue testing it. ;)
> See ya.
>

Hi David.

>Thanks for the backtraces :) (how you made ktechlab crash is also useful=
,
>even if you can't reproduce the crash...)

Well, I was playing with the circuit editor. I was using a potentiometer,=

a transistor, resistors, a battery, a voltage probe and probably a voltag=
e
meter.
I was adding/deleting wires, pausing/resuming the simulation, etc...
I reproduced KTechLab_bt2.txt in a ocassion, but I don't know how to repr=
oduce
it again. The last time that it produced that backtrace I had a circuit i=
n
the screen and then I added a diode (dragging to the board), then the pro=
gram
crashed in the same moment that I released the mouse button.

In one ocassion I got the following warnings in the console:
ktechlab: WARNING: ConRouter::splitConnectorPoints: Could not find point
(180,276) in connector points
ktechlab: WARNING: ConRouter::splitConnectorPoints: Returning generic lis=
t

>I've hopefully fixed two of the bugs you sent me (and discovered a new o=
ne
>while investigating...), but I'm still unsure about what's going on in
>KTechLab_bt2.txt - some sleep should help here (same applies to
>potentiometer glitch).

Well, I you had enough, here's another backtrace with another fascinating=

undescovered bug waiting you to fixing it. ;)

Another backtrace. ;)

On Friday 21 October 2005 10:29, deifo@... wrote:
> Another backtrace. ;)

Thanks :) 

Backtraces 2 and 5 seem to be the same issue, which I still can't work out why 
it's happening...

Backtraces 1 and 4 are both caused by similar problems in the Simulator class, 
which should all be fixed now (there were quite a lot of other potential 
crashes as well) - so you should be getting less issues there :)

And if you've been updating from svn, the misaligned pin in the rotated 
potentiometer (and other components) are also fixed.

> In one ocassion I got the following warnings in the console:
> ktechlab: WARNING: ConRouter::splitConnectorPoints: Could not find point
> (180,276) in connector points
> ktechlab: WARNING: ConRouter::splitConnectorPoints: Returning generic list

Nothing too serious there, although it's still a bug that needs fixing. This 
error just occurs when creating a T-junction of wires, and the new wire 
didn't have its end in quite the right place.

Another crash while rightclicking.

Now the program seems far more stable. I tried to reproduce backtraces 1
and 5 with no luck.

Also these messages appear in the console when I execute "undo".

ktechlab: ERROR: [Node::Node(ICNDocument*, Node::node_type, Node::node_di=
r,
const QPoint&, QString*)] Could not register id node1
ktechlab: ERROR: [Node::Node(ICNDocument*, Node::node_type, Node::node_di=
r,
const QPoint&, QString*)] Could not register id node1__20
ktechlab: ERROR: [Node::Node(ICNDocument*, Node::node_type, Node::node_di=
r,
const QPoint&, QString*)] Could not register id node1__24
ktechlab: ERROR: [Node::Node(ICNDocument*, Node::node_type, Node::node_di=
r,
const QPoint&, QString*)] Could not register id node1__28

On Friday 21 October 2005 16:14, deifo@... wrote:
> Another crash while rightclicking.

Hmm...this crash looks possibly like a bug in kde/qt libraries - there is 
nothing special about the last line in the backtrace that is part of ktechlab 
sources. I'll keep a lookout, but I've never managed to crash ktechlab by 
right clicking yet...

>
> Now the program seems far more stable. I tried to reproduce backtraces 1
> and 5 with no luck.

Good to hear :)

>
> Also these messages appear in the console when I execute "undo".
>
> ktechlab: ERROR: [Node::Node(ICNDocument*, Node::node_type, Node::node_dir,
> const QPoint&, QString*)] Could not register id node1
...
>
> ktechlab: WARNING: ConRouter::splitConnectorPoints: Could not find point
> (180,276) in connector points

As of revision 508, both these messages should no longer occur :)

Hello.

I was testing Ktechlab from yesterday SVN and it crashed 2 times just cli=
cking
on the screen. The backtraces are attached.

Thank you for your effort.

Thanks for the backtraces.

ktechlab-bt1.txt - I've seen this crash a few times before, and have so far 
been unable to find a solution. It seems like it could be a bug with KDE's 
libraries - so I'll replace the code in use (which needs replacing anyway), 
and that should fix it.

ktechlab-bt2.txt - fixed (the bug was introduced recently when adding support 
for european-style logic gates).

David

On Saturday 28 January 2006 14:58, deifo@... wrote:
> Hello.
>
> I was testing Ktechlab from yesterday SVN and it crashed 2 times just
> clicking on the screen. The backtraces are attached.
>
> Thank you for your effort.