OpenDrain not working

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

OpenDrain not working

Klaus Rudolph
Is there still anyone????

After years of absence I try again to run the current git HEAD. As known
OpenDrain and LCD is not working in that version.

I try to get OpenDrain running but I run against a wall. Can anyone help
to get that working again?

There are so many changes between that last working version and the
current head that I have no idea what finally breaks that feature.

Regards
  Klaus

Reply | Threaded
Open this post in threaded view
|

Re: OpenDrain not working

Erik Christiansen
On 27.10.19 15:09, Klaus Rudolph wrote:
> Is there still anyone????

Yes, but no response to queries or patches for some years, so I figure
we're on our own now.

> After years of absence I try again to run the current git HEAD. As known
> OpenDrain and LCD is not working in that version.
>
> I try to get OpenDrain running but I run against a wall. Can anyone help
> to get that working again?
>
> There are so many changes between that last working version and the
> current head that I have no idea what finally breaks that feature.

The last working version is perhaps the best one to move forward, then.
If any subsequent change is mission critical, a cross-branch merge of
just that should be fairly simple. (At least it is in cvs, which I still
use at home.) Trying to diagnose and fix abandonware is not a survival
strategy, I figure.

It is a long time since I last used simulavr, as diagnostic logging on a
running AVR usually suffices for my needs.

Isn't the shiny new "It has to be C++" version of simulavr your baby?

Erik

Reply | Threaded
Open this post in threaded view
|

Re: OpenDrain not working

Klaus
Hi,
> Yes, but no response to queries or patches for some years, so I figure
> we're on our own now.

Great! :-)



>
>> After years of absence I try again to run the current git HEAD. As known
>> OpenDrain and LCD is not working in that version.
>>
>> I try to get OpenDrain running but I run against a wall. Can anyone help
>> to get that working again?
>>
>> There are so many changes between that last working version and the
>> current head that I have no idea what finally breaks that feature.
>
> The last working version is perhaps the best one to move forward, then.
> If any subsequent change is mission critical, a cross-branch merge of
> just that should be fairly simple. (At least it is in cvs, which I still
> use at home.) Trying to diagnose and fix abandonware is not a survival
> strategy, I figure.
>
> It is a long time since I last used simulavr, as diagnostic logging on a
> running AVR usually suffices for my needs.
>
> Isn't the shiny new "It has to be C++" version of simulavr your baby?

Yes, I had started that migration to C++. And after a while a lot of
people bring a lot of new features and add a lot of good stuff. But some
things are broken now...

I spend some hours now to see what is the problem on the OpenDrain and I
got a solution. But there is definitely a bit more of unit testing todo
to let that be valid change...

The changed lcd implementation is incompatible with my "older" one.
maybe I will add simply a second one "lcd2" or so.

The tracer stuff is quite complicated and I still have no idea what and
why is going on. But it should be possible to hack some more infos to
the trace. What I am really missing is all the values. In "old" version
we have the content of each memory and register change. Current version
only prints out the assembler command and nobody knows what the
instruction has done.

But before adding somethings, the prog crash should be fixed. The tracer
crashes after a few lines, no idea if that is a memory corruption or
buffer overrun or.... simply not looked into it. But it only needs
seconds to crash. I wonder why nobody has seen that as it is in there
for a long time. Seems really nobody is using the tool :-)


Maybe avr is totally outdated and all the people are running arm
devices. Don't know. :-)

OK, if you or someone else has some fun on fixing.... :-)

Regards
  Klaus





>
> Erik
>

Reply | Threaded
Open this post in threaded view
|

Re: OpenDrain not working

Thomas K
In reply to this post by Klaus Rudolph
Hi Klaus,

I assume, that you have checked the HEAD of master branch?


I don't know, what you have tried and what's wrong. But the Pin
behaviour and also the OpenDrain object is unchanged and working. Please
see the regress/gtest/session_io_pin test case (it's from you!) This
test is working. Maybe your case today is different, then please
describe it or maybe write a new test. If this is connected with the tcl
part, then we have to think about tests on tcl.


For me it's a question of time to have the possibility to check or fix
it, so any help is welcome. In the moment I have made a try to change
from automake tools to cmake, but it's incomplete in the moment (core is
working, but not all cases) And included, to transform the python
regression test to python3.


And I'm working to create a docker image for simulavr to be able to
remove the windows code parts but to be able to use simulavr also on
windows (win7 or also win10) without a pain or some fixes especially for
this windows ports.


cu, Thomas


Am 27.10.19 um 15:09 schrieb Klaus Rudolph:

> Is there still anyone????
>
> After years of absence I try again to run the current git HEAD. As known
> OpenDrain and LCD is not working in that version.
>
> I try to get OpenDrain running but I run against a wall. Can anyone help
> to get that working again?
>
> There are so many changes between that last working version and the
> current head that I have no idea what finally breaks that feature.
>
> Regards
>  Klaus
>