Improved Gates & Triggers Script, Overhauled screensaver

[chrisib]

New Script: Gates & Triggers

Implementation of the modified G&T script as described in #307 (comment). din or b1 is an incoming gate or trigger signal that is processed as follows:

• cv1 outputs a gate whose rising edge corresponds with the rising edge on din. The length of this gate is determined by k1, k2, and ain
• cv2 outputs a 10ms, 5V trigger on the rising edge of cv1/din/b1 (since those all share the same rising edge)
• cv3 outputs a 10ms, 5V trigger on the falling edge of din/b1
• cv4 outputs a 10ms, 5V trigger on the falling edge of cv1
• cv5 outputs a gate whose state toggles every time a rising edge is detected on din/b1/b2
• cv6 outputs a 10ms, 5V trigger on the falling edge of cv5

Screensaver Overhaul

Added the new experimental.screensaver.OledWithScreensaver class, which acts as a 1:1 replacement for europi.oled but with automatic support for activating a screensaver or blanking the screen.

Whenever the user presses a button, turns a knob, or otherwise interacts with the module in a way that should clear the screensaver, the user should call the notify_user_interaction() method to reset the timer. Otherwise after 5 minutes of idle time, the .show() method will show the screensaver instead of whatever is in the oled frame buffer. After 20 minutes of idle time, the screen will go blank.

As part of making this work as a 1:1 replacement the centre_text method has been modified; it no longer automatically calls self.show() at the end, but the old behaviour can be re-enabled by passing the auto_show=True argument to the method. All existing calls to centre_text have been modified accordingly to ensure no regressions.

The new OledWithScreensaver class is used in Pams, Logic, Euclid, and G&T.

[mjaskula]

Hey @chrisib, thanks for all of the work you are doing. Any chance that you can slim this PR down a bit, or possibly spit it up? It has two main ideas (the new script and the screensaver changes). And then on top of that there are several files with changes that don't have anything to do with either of these ideas.

Lately it's been hard for me to find enough time to give these PRs the attention that they deserve, and it seems that the same is true for the other reviewers. I apologize for that. Anything that you can do to reduce the size of your PRs will go a long way to help.

In the meantime, I'll try to work through this as is, but it's going to take me a few sessions

[chrisib]

Because all the changes were made in branches off of branches I think the amount of work it would be to de-couple the two ideas may wind up being more than I can necessarily take on right now. The new script was written with the screensaver changes already made. I could move the new script out into its own MR, but we'd still have a situation like #311 where the screensaver changes are present anyway. (Assuming this MR gets approved I plan to rebase #311 and update that MR to reduce it.)

I know this is a pretty big MR, so there's definitely no expectation on my end for it to get reviewed & merged right away. Definitely take your time.

I will go ahead and revert a few files that only git linting changes; it looks like while I was applying the new screensaver changes & modifying the centre_text function a few files got trailing whitespace & other cosmetic changes applied by my IDE. I'll revert those changes at least to keep this to just my functional changes

[mjaskula]

I've looked over the new script in isolation and it looks pretty good. Just a couple of comments. I'll look over the screensaver in my next session.

[chrisib]

I've just reverted the changes I made to some of the other scripts (pams, logic, etc...). That should narrow the scope of this MR. I'll make a new branch and re-apply those changes there.

Edit:
Screensaver changes to Pam's, Logic, and Euclid have been moved here: #312

That should reduce the scope of what you need to look at for this MR. Thanks!

[roryjamesallen]

I think I'm caught up with the changes this PR is making, but let me know if I'm clearly missing something already discussed:

I can't quite work out the benefit of a screensaver that's only based on time, especially as this also adds the complexity of users needing to check for user input and manually call the notify_user_interaction function.
In my mind, a more useful screensaver would test the current frame buffer against the new one each time oled.show() is called, and if the buffer is the same and 5 minutes has passed, then enter screensaver mode.

This would also fix my other issue with the screensaver, which is that some scripts won't get touched but would like to be seen - I often leave harmonic LFOs running, not touching any controls, but producing a cute display in the corner of my rack, and this implementation would turn it into a screensaver.

I think maybe I'm misunderstanding how programmers of new scripts can implement this screensaver and whether once moved from experimental to main, it'll be 'compulsory'. I know that in a lot of my scripts I'd rather risk a little bit of screen burn-in than have to touch a control to wake up the screen, because often touching a control will also alter the actual script, without there being a dedicated 'wake screen' button

[chrisib]

My thinking was that the screensaver-supported class would exist alongside the existing one, not replace it completely. Several scripts (mostly ones I've written) have duplicate code & logic to provide a basic screensaver (Pam's, Logic, G&T, Euclid, Quantizer, Seq. Switch). This MR is mainly to consolidate those and cut down on code duplication.

For scripts that have a continuous, useful display (e.g. Harmonic LFOs, Particle Physics) the old Oled class can be used since there's no point in them having a screensaver. But for scripts where burn-in is a potential hazard, the new class exists as an option, not as a requirement. The goal is absolutely not to make this class mandatory for all future scripts, but rather make it a convenient parallel to the lower-level, no-screensaver europi.oled member.

New or existing scripts wanting to use the screensaver would simply need to:

1. create an instance of the new OledWithScreensaver class
2. replace all instances of oled.* (e.g. .show()) with the similarly-named member from OledWithScreensaver
3. add a call to notify_user_interaction inside any functions or ISRs that should wake the the screen up if the screensaver is active

Most real-world screensavers I've run into are time-based, with user interactions waking them up/resetting the timer before they cut in. Yes, checking the underlying frame buffers for pixel differences instead of a static timer could work, but that's a lot of CPU overhead.

[mjaskula]

I just started my review over, since it had been a while.

I only had a few minor suggestions to point out, but over all this looks pretty good.

[chrisib]

I just sync'd my fork with the latest main & rebased this branch, and the CI seems to be failing now. I don't have time to look into it immediately, but it looks like an issue with the bootloader menu.

[chrisib]

Looks like it was just something weird in the rebase. I copied europi.py directly from the latest main and it seems to be working again.

[mjaskula]

This looks good to me now. Good work.