M5Core2 Library 0.0.3 Issue

[helo-head]

Does not compile with Version 0.0.3 of the M5Stack M5Core2 Library ... You must currently use 0.0.2

[brianteeman]

as the current version is 0.1.4 I assume this is no longer relevant?

[helo-head]

I confess I have not tried to recompile it since last year due to all the M5Stack library issues. If you find an a set of libraries that work reliably let me know and we can update this comment /and/or readme and add any specific instructions needed to get it to compile.

[brianteeman]

yeah - so far no luck at all getting this code to compile. I'm very new to this (but not development)

[brianteeman]

So I found the "problem". Basically you need to be using Button instead of TouchButton

I was in the process of starting to convert everything when I stumbled across https://github.com/kitazaki/M5Core2_atemswitcher

This is a fork of @aaronpk code to work with the newer m5 libraries. It said it was only tested with 0.0.6 but I just compiled it without error on 0.1.5 so I assume the lbraries have not changed for this part at least since then.

Not tested it on the actual atem yet as that's in another building but the fact it compiles gives me hope.

[helo-head]

As you figured out, the TouchButton type was dropped from the M5Core2 Touch library in 0.0.3 with no documentation other than a single note on GitHub in the M5Touch.h header that refers to a Legacy API. Unfortunately even the header incorrectly states the "This older API is still supported". Many thanks for opening a bug against M5Core2. These types of issues are why I stopped updating the code last year.

The good news I was able to validate the code does compile and work using:

Arduino IDE 1.18.13 / M5Stack Board 1.0.7 / M5Core2 Library 0.0.2

Of course I understand this is not a desirable outcome, as such I will take a closer look and see what it will take to update the code to the new Touch library.

[brianteeman]

Not tested it on the actual atem yet as that's in another building but the fact it compiles gives me hope.

Tested that one now and as soon as I remembered to modify the atem libraries as they stated it worked

[brianteeman]

Progress so far in converting your script. https://gist.github.com/brianteeman/30447842f7b3c07a091a753cc57a6a36

now to look at the sd card code so I can actually test this

[helo-head]

Ah, looks like we started working on this at the same time. Im finished with the required changes however I have run into an issue with "I2S Port 0 not installed". As my Core2 was previously used for testing it still had a config stored in NVM so at least Im not dead in the water. Now I just need to borrow an ATEM for testing as well as try and figure out whats going on the I2S system.

[brianteeman]

sounds like it could be an issue with the drivers on your computer and not the Core2.

Would be good to see your code as I am working kind of blind and guessing what it is supposed to do whereas you know what it is supposed to do as you wrote it.

[helo-head]

Try the attached, In theory it should work. I had to add a quick hack to the readSDconfig function since cfg.begin apparently now requires the full path even if the file resides in the root. If it works let me know and Ill start cleaning it up before I upload it.

Tallly-Core2-V10B.txt

[helo-head]

At some point past IDE 1.8.13 the SD File name() or possibly openNextFile() function was changed so that the name returned no longer includes the full path when dealing with the root directory This is why the SDConfigFile.begin() code was failing on the subsequent call to SD.open. Apparently SD.open always required the full path which should not be case when dealing with the root directory. I would not have figured this out had I not taken a look at the screen shots from version 1.0 that showed the fully qualified names.

Ill ping the developer and see what he thinks about both issues . In the mean time my hack seems to work as a quick fix until I can determine the best way to correctly address this issue.

[brianteeman]

didnt have a chance to look at anything today. But I brought the atem to my home office so will check out your script tomorrow.

[brianteeman]

Great news. It works.

Now I can start customising it for my own purposes. The first step of which is to remove alll the very clever config stuff you wrote and replace it with hard coded values. I understand why you did it that way but it is overkill for my usecase and it requires so much code to maintain

[helo-head]

I started the same way, however within a month of the first ATEM purchase the company opted to expand to several thus the reason for the SD card allowing easy reconfiguration. However leaving SD cards around with sensitive info seemed inappropriate thus the reason for writing the config data in flash allowing us to keep the SD cards in a secure location until needed for a reconfiguration. At last check they expanded the fleet even more with some of the newer models.

[brianteeman]

yes I changed my mind.

such a shame that the skaarhoj library stopped at 7.5 and that the replacement opensource library (libatem) is in c# not c++