ProffieOS v6.x ALPHA testing

YaMateJack · October 22, 2021, 8:47am

I’m getting “error in font directory” when I boot up my saber and when I change fonts. All I have on the SD card is Kyberphonics The return and fett 263’s common folder. I have only added tracks and a push folder to the font folder. Any ideas?

Fett263 · October 22, 2021, 11:43am

Check none of your sounds are out of sequence or skipping numbers.

Example: blst01, blst02, blst03 not blst01, blst03, blst04

YaMateJack · October 22, 2021, 1:30pm

I checked every sound, and nothing skipped. Nothing had changed. I then swapped everything back to default, the same way I got it from both Fett263 and Kyberphonic. I still got the same message.

YaMateJack · October 22, 2021, 1:32pm

Apologies, I wrote that as if I was not talking to the guy I was writing about, completely slipped my mind XD

Fett263 · October 22, 2021, 1:48pm

If it’s saying “error in font directory” on boot then it’s not something unique to OS6, it means there’s an error in the font directory. You can confirm by testing the font as-is using OS5.9. If the font works in OS5.9 (without changing it) then that’s really odd, if it still says “error in font directory” then there’s an error in the font directory and you just need to find it.

I believe you can try typing “effects” in Serial Monitor and it may print out which have an error but it’s been forever since I did so it may be a different command. But on boot there’s nothing unique to OS6 that would cause that error message that I can think of, so rule out the font having an error by trying it on OS5.9 first, then we’ll go from there.

Afrojedi · October 22, 2021, 1:50pm

I had a similar issue make sure that you don’t have any fontname01(copy).wav I had accidentally copied files to my sd card and instead of overwritting it had renamed all the files. delete the copies from the sd card.

was this in reference to using the alpha testing of os6?

profezzorn · October 22, 2021, 2:18pm

Try the “effects” command in the serial monitor. It will tell you which effect has a problem.

NoSloppy · October 22, 2021, 9:33pm

I think “effects” is restricted if you have #define DISABLE_DIAGNOSTIC_COMMANDS in your config. Check that if you just get “Whut??” in serial monitor.

NoSloppy · October 23, 2021, 1:06pm

I have this mostly working as is right now. USBD_Connected works 66% - good while still connected from upload and responds to disconnection. But when re-plugging USB, while the OLED responds by stopping idle.bmp, the SD doesn’t mount. That might just be my computer? So some other testers might be helpful.
Also the idle.bmp is alternating each loop with battery meter and repeats.

profezzorn · October 23, 2021, 7:40pm

I’m not sure what happens if ProffieOS is not fast enough to react when USB is plugged in. The computer might decide that the drive busy before ProffieOS gets around to unmount the SD card.

Ideally, the computer will try again later and everything will just work out, but I’m not sure if that’s actually what happens or not.

NoSloppy · October 23, 2021, 9:21pm

Maybe I’ll try playing with this, although what does the delay protect from?

github.com

profezzorn/ProffieOS/blob/5f8bfe4a808c7a1fa356e45b0a836840c54f5bc0/common/sd_card.h#L26

    
      
              if (amplifier.Active() || AudioStreamWork::sd_is_locked()) {
                last_enabled_ = millis();
                return true;
              }
          #endif    
              if (SaberBase::IsOn()) {
                last_enabled_ = millis();
                return true;
              }
              uint32_t t = millis() - last_enabled_;
              if (t < 1000) return true;
              return false;
            }
          
          
  void Mount() {
              last_enabled_ = millis();
              if (LSFS::IsMounted()) return;
              // Wait for card to become available, up to 1000ms
              uint32_t start = millis();
              while (!LSFS::CanMount() && millis() - start < 1000)
                armv7m_core_yield();

I also have a PC and a virtual machine I can test with tomorrow, but in the meanwhile, anybody with a Windows machine and an OLED care to take a crack testing this?

github.com

profezzorn/ProffieOS/blob/0d53e939e65de96e39cd96313e172028b59df373/display/ssd1306.h

#ifndef DISPLAY_SSD1306_H
#define DISPLAY_SSD1306_H

#include "monoframe.h"

// Images/animations
IMAGE_FILESET(font);
IMAGE_FILESET(boot);
IMAGE_FILESET(on);
IMAGE_FILESET(clsh);
IMAGE_FILESET(blst);
IMAGE_FILESET(lock);
IMAGE_FILESET(force);
IMAGE_FILESET(idle);

template<int Width, class col_t>
class Display : public MonoFrame<Width, col_t> {
public:
  virtual void Page() = 0;
  virtual void SB_Top() = 0;

This file has been truncated. show original

@SA22C ? @Afrojedi

SA22C · October 24, 2021, 7:10am

This is an odd one, but I’ve been testing OS6 with @Fett263 to help with his 1 button merge into his prop file. On two different sabers, I’ve seen some very strange behaviour:

When connected to USB, sometimes after plugging the connector, the saber will randomly lose all motion. Clash, swing and hum don’t play and no gestures work. There are no errors in the log when viewing the serial monitor: it just doesn’t register any movement or acceleration, and cutting power to the saber ‘fixes’ the problem until the next random occurrence.

I am failure sure the issue is confined to times when the saber is connected via USB, but I can’t say for sure until I do more testing.

Edit to add:

Flashing back to 5.9 seems to solve the problem.

The sabers are both proffieboard v2 boards but from different sources: one is an Artekit and the other is a LGT board. I’ve checked wiring on both and see no evidence of solder bridges or wire strands grounding out. Neither board overheats or shows any sign of physical issues.

Afrojedi · October 24, 2021, 7:30am

sorry dude, no oled and linux based.

profezzorn · October 24, 2021, 9:03am

Some parts of the motion code has been rewritten in OS6, so it’s not entirely surprising if something doesn’t work.

Now for the questions:

I don’t suppose you see “motion disable” messages in the serial monitor?
Does an “on” command make it work again?
Do you have anything else on the i2c bus? (say an oled display?)
Does this happen while the saber is ignited, or when it is off?
If you post your config, I can try to see if I experience the same problem with it.

Finally; Unless a solution is found by tomorrow, I will try adding some debugging commands that dump some i2c state. If none of the other things above helps, maybe that will tell us something.

SA22C · October 24, 2021, 2:28pm

no ‘Motion Disable’ or other error messages of any kind
haven’t tried that, once I see the error again I’ll try
Nothing. Both sabers are simple installs
It seems to happen when the saber is off, because it crops up when switching between presets, not while the saber is ignited

Config attachedmhs_1btn.h (127.6 KB)

profezzorn · October 24, 2021, 6:45pm

Ok, I’ve improved the “i2cstate” command a bit.
You’ll need ENABLE_DEVELOPER_COMMANDS to be able to use it though.
Once it ends up in this weird state, run “i2cstate” in the serial monitor, and it should tell some things about what’s going on.

When this happens, how are you actually switching presets? (Or does it matter?)
(I’m particularly curious if I should expect it when doing “n” in the serial monitor, or just when using buttons to do it.)

profezzorn · October 24, 2021, 6:47pm

Oh, please also run the “top” command to see what it says.

profezzorn · October 24, 2021, 9:46pm

One more thing @SA22C, what version of the arduino-proffieboard plugin area you using?

SA22C · October 25, 2021, 9:57pm

I’m using 3.6 Experimental for the plugin.

As is the usual with these kinds of things: had the problem occur four times in a night before I posted, now it hasn’t happened at all.

profezzorn · October 25, 2021, 10:24pm

If we can’t reproduce the problem anymore, we’ll have to revisit it if it pops up again.
Maybe when we go to BETA, we’ll get more users and more reports.