Open PLC Project/community

Pete 'mash# Morgan

@pedromorgan_gitlab

This is part of my cunning plan to create the "cookbook" site which will link all the stuff together. (my dream)

we just setup redirects on the links.. no problem there..

the sitemap.xml etc will help there big time..

But this is the problem i forsee. .and where trying to help with eg the firmware..

So atmo the main openplcproject.com links to downloads and zips

instead the repos is here https://gitlab.com/openplcproject/openplc_firmware

and using gitlab knocks out the tarballs.. etc https://openplcproject.gitlab.io/openplc_firmware/platforms_index.html

Pete 'mash# Morgan

@pedromorgan_gitlab

https://openplcproject.gitlab.io/openplc_firmware/arduino_mega_pinouts.html

Garret Fick

@garretfick

It will eventually get there, but I'm focused on the minimum set of things that I need to do to push out a release before the end of the year.

From that perspective, I'm only focusing my time on two things:

integration testing
updated docs for the changes that have been made

Pete 'mash# Morgan

@pedromorgan_gitlab

2 is here https://openplcproject.gitlab.io/openplc_v3/

Garret Fick

@garretfick

That's part of it, but not all. I'm updating the user (not API) docs right now

Garret Fick

@garretfick

Just want to put out what I'm proposing for integration tests
https://github.com/garretfick/OpenPLC_v3/tree/feature/INTEGRATION_TEST

I need to get modbus tests implemented first, but it is coming soon

Feedback is appreciated

Thiago Alves

@thiagoralves

Where is your user documentation for the changes?

Is it in your repository?

Garret Fick

@garretfick

Nope, for more it is just in the changed files

But it is quite small

Garret Fick

@garretfick

@thiagoralves Have you ever checked if the response from write coils is correct? For some reason, it is always coming back true. However, it is correctly interpreting the write message and the read response updates based on the write message?

# Set the coil to true
resp = client.write_coil(0, True, unit=unit)
self.assertEqual(1, resp.value)
time.sleep(1)

# Validate that the coil now reports true
rr = client.read_coils(0, 1, unit=unit)
bit_value = rr.getBit(0)
self.assertTrue(bit_value)

# Write the value false to the coil
client.write_coil(0, False, unit=unit)
# THIS FAILS??? self.assertEqual(0, resp.value)
time.sleep(1)

# Validate that the coil now reports false
rr = client.read_coils(0, 1, unit=unit)
bit_value = rr.getBit(0)
self.assertFalse(bit_value)

That's my very simple test - it is very strange that the marked part is failing - I'm almost thinking it is a pymodbus error, but that's hard to believe

Thiago Alves

@thiagoralves

It depends on what your ladder program is doing

Garret Fick

@garretfick

It never sets that value

Thiago Alves

@thiagoralves

PLC logic always superseed modbus requests

Garret Fick

@garretfick

Next step is probably wireshark

Thiago Alves

@thiagoralves

Are you sure? That would be the only reason why it would get stuck on true

Garret Fick

@garretfick

It is very strange, but it isn't stuck on true. The value does change

It is just the response message from the write request that is somehow not right

Thiago Alves

@thiagoralves

Weird

Garret Fick

@garretfick

Yeah, it is very strange

Thiago Alves

@thiagoralves

Try downloading radzio modbus (a free modbus poll program)

Garret Fick

@garretfick

Because the old (and new) code both just don't touch those parts of the message. It just echos back the request

Thiago Alves

@thiagoralves

Then you can check the status of the coil in real time

Garret Fick

@garretfick

No worries... I'll capture the packets

Thiago Alves

@thiagoralves

Also, check the monitoring page on the runtime to see if the coil is changing

Garret Fick

@garretfick

I added debug messages in the runtime, and that part suggest all is ok

thanks :)

Thiago Alves

@thiagoralves

By old code do you mean the code on the master branch?

It only echoes the beginning of the request, the actual data is gotten from the buffers

Garret Fick

@garretfick

Woo hoo! Integration tests are ready for a pull request. They don't cover all capabilities yet, but I'm working on it
thiagoralves/OpenPLC_v3#118

I did discover (and fix) multiple issues where capabilies had diverged

The current tests are implemented in one massive file. I'll be splitting it up in a follow up change as I'm next focusing on the modbus master

Thiago Alves

@thiagoralves

Wow... you never rest?

Garret Fick

@garretfick

Nope. I even managed to have two days of skiing Fri and Sat

Garret Fick

@garretfick

For the sake of general interest, that is my current roadmap that fits with some of the work I've been doing

Key things you might see, beyond features, are being able to put together an "official release" every 4 months. What that means isn't defined, but the basic idea is to develop capabilities that enable more continuous delivery capabilities (even if only released periodically)

Garret Fick

@garretfick

@pedromorgan_gitlab can you point me to the slice code for your website changes?

Garret Fick

@garretfick

I think I found it... had a local copy from a long time ago

Pete 'mash# Morgan

@pedromorgan_gitlab

@garretfick its in the www refactor thiagoralves/OpenPLC_v3#79

Where communities thrive

People

Repo info

Activity