Discard only what's in _fifo, don't wait for more to arrive.

[FozzTexx]

This should fix the problem with NetSIO.

[markjfisher]

Still failing on both config (v1.5.1) and config-ng against fujinet-pc.
Same symptoms as yesterday but in non-esp32; it's looping on same command many times.

From logs trying to boot config v1.5.1.
Note commit id in startup

Startup:

07:59:17.262704 > --~--~--~--
07:59:17.262705 > FujiNet v1.5.1+git-841e3d30 2025-11-04 03:38:22 (ATARI) Started @ 0
07:59:17.262709 > Detected Hardware Version: fujinet-pc
07:59:17.262712 > SPIFFS mounted.
07:59:17.262714 > SD mounted (directory "SD").
07:59:17.262716 > fnConfig::load "fnconfig.ini"
07:59:17.262725 > fnConfig::load read 1124 bytes from config file
07:59:17.262744 > Reading wifi section
07:59:17.262763 > disk UNMOUNT
07:59:17.262766 > FileSystemSPIFFS::filehandler_open /autorun.atr rb
07:59:17.262769 > disk MOUNT
07:59:17.262770 > ATR MOUNT
07:59:17.262774 > mounted ATR: paragraphs=5760, sect_size=128, sect_count=720, disk_size=0
07:59:17.262783 > PCLINK[1] MOUNT "SD"
07:59:17.262785 > Creating a default printer using FS_SDFAT storage and type 1
07:59:17.262791 > SIO SETUP
07:59:17.262794 > NetSIO: No WiFi!
07:59:17.262795 > Suspending NetSIO for 400 ms
07:59:17.262796 > Set HSIO baud from 19200 to 57600 (index 8)
07:59:17.262798 > Setup complete @ 0 (0ms)
07:59:17.262799 > DummyWiFiManager::start
07:59:17.262803 > Starting web server http://0.0.0.0:8000
07:59:17.262869 > DummyWiFi connect to SSID "Dummy Cafe"
07:59:17.662000 > Setting up NetSIO (localhost:9997)
07:59:17.662020 > Resolving hostname "localhost"
07:59:17.662527 > Resolved to address 127.0.0.1
07:59:17.662631 > NetSIO ping localhost time=0.073 ms
07:59:17.662639 > ### NetSIO initialized ###
07:59:17.662640 > NetSIO set_baudrate: 19200
07:59:20.186897 > NetSIO peer baudrate: 127840
07:59:21.915972 > NetSIO peer baudrate: 19040

Then booting config, it loops on doing WIFI ENABLED check

07:59:21.916125 >
07:59:21.916127 > CF: 70 ea 00 00 5b
07:59:21.916688 > sioFuji::sio_process() called, baud: 19200
07:59:21.917012 > ACK!
07:59:21.917014 > Fuji cmd: GET WIFI ENABLED: 1
07:59:21.917016 > ->SIO write 1 bytes
07:59:21.917324 > COMPLETE!
07:59:21.917379 > SIO CMD processed in 2 ms
07:59:22.059379 >
07:59:22.059386 > CF: 70 ea 00 00 5b
07:59:22.059944 > sioFuji::sio_process() called, baud: 19200
07:59:22.060259 > ACK!
07:59:22.060260 > Fuji cmd: GET WIFI ENABLED: 1
07:59:22.060261 > ->SIO write 1 bytes
07:59:22.060569 > COMPLETE!
07:59:22.060612 > SIO CMD processed in 1 ms
07:59:22.164207 >
07:59:22.164215 > CF: 70 ea 00 00 5b
07:59:22.164771 > sioFuji::sio_process() called, baud: 19200
07:59:22.165084 > ACK!
07:59:22.165086 > Fuji cmd: GET WIFI ENABLED: 1
07:59:22.165087 > ->SIO write 1 bytes
07:59:22.165427 > COMPLETE!
07:59:22.165435 > SIO CMD processed in 1 ms
07:59:22.307080 >
07:59:22.307082 > CF: 70 ea 00 00 5b
... etc

[FozzTexx]

Strange, it comes up ok in the emulator I've been testing in.

[FozzTexx]

Can you try this branch and see if it works? If not, can you bisect it and see which commit it breaks at? I started at 31f6107 and kept making incremental commits.

https://github.com/FozzTexx/fujinet-platformio/tree/netsio-siocom-1-class

[markjfisher]

The failing commit is 46be02c:

fujinet-firmware-master on  HEAD (70a76ed) (BISECTING) via △ v4.1.2 via 🐍 v3.13.7
❯ git bisect good
46be02c15099d0c88a12f517470de958e5fd7f5f is the first bad commit
commit 46be02c15099d0c88a12f517470de958e5fd7f5f
Author: Chris Osborn <[email protected]>
Date:   Mon Nov 3 19:27:44 2025 -0800

    Inherits from IOChannel

 lib/bus/sio/NetSIO.cpp | 30 +++++++++++++++++++++++++++++-
 lib/bus/sio/NetSIO.h   | 18 ++++++++++++++++--
 2 files changed, 45 insertions(+), 3 deletions(-)

[markjfisher]

I'm using Altirra with the python bridge which has always been rock solid.
I think atari800 has a different implementation of the NetSIO built into application.

[FozzTexx]

I'm using Altirra with the python bridge which has always been rock solid. I think atari800 has a different implementation of the NetSIO built into application.

Guess I'm spending the day trying to figure out how to run Altirra and reproduce. 😅

[markjfisher]

I'm using Altirra with the python bridge which has always been rock solid. I think atari800 has a different implementation of the NetSIO built into application.

Guess I'm spending the day trying to figure out how to run Altirra and reproduce. 😅

not sure if your distro has it in (arch AUR does).
I run this on my linux box:

altirra config-v1.5.1.com

The bridge is run with:

cd fujinet-emulator-bridge/fujinet-bridge
python3 -m netsiohub

Obviously you have to run a fujinet-pc too.

You'll need to setup your ROMs in Altirra.
In System/Configure System, Computer/Firmware, click the "Firmware Manager..." button

These are the high level ROMs I have installed:

You can dump the ROMs in a folder and add them all in one go, it detects most types. I'll send you a pack.

To configure Altirra to use the bridge:
System/Configure System, Peripherals/Devices
Click "Add..." and right at the bottom of the list is "Other devices/Custom device".
Click that and choose the "netsio.atdevice" from the emulator-bridge/altirra-custom-device folder.
If you're on WINE, you'll have to pick it from the Z drive. My path is:

You will then have a custom device that talks to the bridge:

If the bridge isn't up, it will show an error.

Now you can exit altirra (it should save config to a default place) and then you can rerun your altirra commad again.

If you're running wine directly, you can run something like

wine /path/to/Altirra64.exe

That should be enough to get going.

You can also create/use "portable" config files for altirra and run it with:

wine /path/to/Altirra64.exe /portable /portablealt:my-altirra.ini

and that will create a local file instead of affecting your global one.
It'll create it if it's missing, then use it next time around and update it with any device changes etc.

I find that's a better way to manage it, so you can copy the ini file around.

[FozzTexx]

Got altirra running and I'm able to reproduce the problem at the same commit.

[FozzTexx]

Works for me in Altirra now

[markjfisher]

Same here. Excellent!