[Ktechlab-devel] I googled recent pages with ktechlab

[santiago gonzalez]

2009/10/10 Zoltan Padrah <zoltan.padrah at gmail.com>

> 2009/10/10, santiago gonzalez <santigoro at gmail.com>:
> > Why not doing a relase from ktl-0.3.7?
>
>  Releasing something on sourceforge needs some rights. Currenly nobody
> active on the list has them, as far as I know.
>
Ok... i understand, but just a svn branch working is enought for the people
to compile it and have it running.


>
> >
> > It works good enought for me, just need to fix the flowcode (changing 1
> line
> > in fpnode) and fix some problems in pic simulation (changing 1 line in
> > gpsimprocessor).
>
>  I know it's a huge shame for me, but there are lots of mails in my
> todo folder, for which I haven't got time to do. Exactly what to
> change and where?
>

Well... here are the new lines:

FIXING FLOWCODE:
_____________________________________________________________________________

src/fpnode.cpp ,one line changed and some commented (around line 38)

FlowPart *FPNode::outputFlowPart() const
.........
.......
     if( !m_outputConnector || !m_outputConnector->endNode() )

    //if( m_outputConnector->endNode() == 0)
    //    return 0;
_____________________________________________________________________________

.
FIXING PIC SIMULATION:
_____________________________________________________________________________

src/electronics/gpsimprocessor.cpp , one line changed and some commented
(around line 270):

void GpsimProcessor::executeNext()
.........
........
........

    //if(get_bp().have_interrupt())
    //{
    //    m_pPicProcessor->interrupt();
    //}
    //else
    //{
        m_pPicProcessor->step(1, false); // Don't know what the false is
for, gpsim ignores its value anyway

        // Some instructions take more than one cycle to execute,
        //so ignore next cycle if this was the case

        if ( (get_cycles().get() - beforeExecuteCount) > 1 )
            m_bCanExecuteNextCycle = false;
    //}
______________________________________________________________________________



> > There is also a bug when closing files directly shown in screen (created
> in
> > /temp/k.. ) that can be easily fixed (just don't creating a new temp
> file).
>
>  This is a new one for me. Which mail haven't I read?
>

Sorry i tought that this bug was reported, ktl crashes sometimes when
closing a document.
i did a fix that works for me, but this way when generating new documents
that are not saved to disk (view directly) the last view will close and the
new will open, then you can't have several views of this kind.
This is not a real fix, just a workaround, but ktl don't crases.

_____________________________________________________________________________

src/textview.cpp , some lines commented (at destructor):

TextView::~TextView()
{
    //if ( KTechlab::self() ) {
    //    if ( KXMLGUIFactory * f = m_view->factory() )
    //        f->removeClient( m_view );
    //
    //    KTechlab::self()->addNoRemoveGUIClient( m_view );
    //}
    delete m_pViewIface;
}
_____________________________________________________________________________


>
> >
> > For the piccomponent is not very difficult to add a property to choose
> clock
> > MHz and add a loop in simulator to manage speed.
> >
>
>  This is tricky. Some odd behaviour might occur if we have 2 loops in
> the simulator.
>

Well... this a 5-step loop inside the logic loop, perhaps i'm wrong but it
is working for me with any problems (by now).
Anyway this should be examined for someone that knows ktl in deep.



> > Running the qtimer at 10 ms does the simulation to go at real time for
> me.
> >
>  Changeing the simulator's "tick" period should be done one day. There
> are quite a few issues with the simulator, starting from the total
> lack of documentation about it.
>
> Ok, i know this issue should be improved one day, but personally i prefer
having the simulation at real time today, and this just works for me:

simulator.cpp __________________________________________________

stepTimer->start(10);

const unsigned maxSteps = unsigned(LINEAR_UPDATE_RATE / 100);
______________________________________________________________


> I know there are a lot of people that wants to use Ktechlab for
> educational
> > purposes, but actually they can't find a ktechlab that works properly.
> >
>  For instance me. And after seeing the codebase, I found this project
> too challenging to not to join it. :))
>

Yes... that's why i think is better having a ktl that works today than
having nothing, even when it is not very accurate, but for educational
purposes is enought if ktl doesn't crash and have flowcode and pic
simulation working.


Another cuestion that i made but didn't  comment is replacing Microbe by
GcBasic, that is a good open-source Basic compiler that support most of pics
and have functions libraries... but i think this is out of this discussion.
I did for my personal use and for some friends that want to use ktl in
digital electronics teaching.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/ktechlab-devel/attachments/20091010/066bc771/attachment.html>