Poll drive directory for changes & regen w/o changing directories?

[benjamink]

The live update seems to require changing out of the directory for the live image & then changing back into it for the refresh/update to happen. Is this correct? If so, could logic be added to watch the directory for file changes & regenerate without the need to change out of & into the directory to refresh?

I'm thinking something along the lines of inotify but more native in the Go code.

(Moving from ProDOS-Utilities discussion)

[benjamink]

Related to changes in #122

[tjboldt]

Did you rebuild the apple2driver after pulling the latest from upstream? It sounds like it's still using the old code. You can verify this by -RPI.COMMAND and see if you get 0027 or 0028.

[benjamink]

I thought I had but I did get 0027 back so I must have missed that. I went through the update steps in #65 & am now seeing 0028. But I'm not seeing new files show up in CATALOG output when I create/rename files in the driveimage directory.

I also noticed that after you symlinked the .bas files to the apple2driver directory that you are now dumped into that apple2driver directory when you use -SHELL instead of the driveimage directory. Seems like ending up in the driveimage directory is preferable.

[tjboldt]

The .bas files are symlinked into the the driveimage directory... Not sure why you are seeing them in apple2driver...

https://github.com/tjboldt/Apple2-IO-RPi/tree/main/RaspberryPi/driveimage

Also, while not automatically polling yet, you can manually trigger a drive regen from the Apple II...

]-RPI.COMMAND
]RPI a2drive 1 regen

[benjamink]

As I re-read your response it dawned on me that the BASIC files being symlinked have nothing to do with where -SHELL puts me on the RPi. Still not sure what would have changed so that -SHELL puts me in the apple2driver directory by default now, though.

As for the a2drive command -makes sense. I didn't think the functionality to auto-update was there yet like I was describing above.

[tjboldt]

Because it no longer uses the current working directory to generate, the code that set the initial working directory to driveimage was removed. It could be reinstated easily. I think for most people that makes sense although for me, I had apple2driver by default for easier in-system development.

[benjamink]

Maybe a config file in the future?

[tjboldt]

I updated it to default to the driveimage directory.

[benjamink]

Nice! Updated & confirmed.

[benjamink]

I've noticed something else that I think crept in recently. It seems the issue where it's impossible to do anything after exiting from -SHELL back to BASIC (#58 ). It seems like access to the virtual disk actually stops working as entering Bitsy Bye after -SHELL also doesn't show any disk contents & hangs.

Resetting the Apple2 gets things back to working again.

[benjamink]

Confirmed versions:

Firmware: 000F
RPI Command: 000E
Driver: 0028

[tjboldt]

This is definitely a bug. I doesn't happen every time but it is reproducible. It is different than bug #57 which would completely hang the RPi code. At least now a reboot of the Apple II fixes it due to commit #58. Would you care to open a new bug? If not, I can open one later.

[benjamink]

I can open one later or tomorrow but I'd be happy to. I'm hoping I can spend some time tomorrow troubleshooting it myself too.

[benjamink]

Opened bug issue: #130