matherp said  

V6.03.00RC17
PicoMiteRP2040V6.03.00RC17.zip
PicoMiteRP2350V6.03.00RC17.zip


Fixes bug in MEMORY SHARE and improves detection of incorrect wiring
Fixes bug in PORT command when using pins > GP32 on RP2350B processors
Adds support for additional TLS crypto methods

There still seems to be a problem with beautifying. Lines 100-120 should line up with line 20 since the FOR N=... loop is now finished.

You need to put next n on a separate line. I won't comment on the programming style  

  matherp said  

Try this

PicoMite (2).zip

Sorry Peter,

That did not solve it.

Volhout

  matherp said  

I won't comment on the programming style  

I wouldn't feel bad. It's David Ahl's code!  

@Peter: I have a vague recollection that several months ago you provided an option or command which could vary the allocation of heap vs regular memory, but I can't find it in the manual or in my code or options or by searching posts. Do I imagine it? Once again, I need more heap. Webmite, no PSRam.

(Note: I asked Claude Code if it had introduced any string arrays without LENGTH values. It had, and adding LENGTH fixed the heap problem. But I'd still like to know if there is a way to adjust heap/regular.)

It did for me so I need absolutely explicit instructions as to how to get a RP2040 into the state where it fails. I nuked it, powered off, and then loaded and it ran your quadrature encoder S/W perfectly - I even wired up a real encoder to make sure the test was properly valid

There is no trade-off between heap and "ordinary memory". An amount of memory is allocated for global and local variables, the rest is heap. What you can do is trade between global and local within the total allowed variables
Edited 2026-06-11 03:29 by matherp

Peter,

The most obvious is when you start the program, and the count is not zero.

Volhout

But I can't reproduce that. If I can't get a Pico into that state I can't possibly fix it.

Peter,

I understand. Currently I am trying to try to find a relation between USB (console) and the problem. Repairing it happens by interacting with the pico (either MMCC or Putty) and editing (Either MMEdit or built in Editor). So repairing always involves (for me) with USB / USB console.

So I have a few Pico's set up with autorun enable. I power them every few days. When they show the problem again, I could try to power them from 5V on Vsys (not USB) and see if it repairs, or the problem persists (*).

Waiting for the problem to re-appear.

Volhout

(*) When the problem shows, a cold boot (pulling/re-inserting USB power) does NOT repair. Let's see if repower without connecting USB repairs.

P.S. I see 2 explanations for this: #1=non initialized RAM, #2=overlapping buffers. #1 can be diagnozed by wiping all RAM at earliest start point with 0, or &hffffffff, before starting anything. Even before USB. #2 is what I am trying to analyze now.
Edited 2026-06-11 16:56 by Volhout

You can try this version and see if it makes any difference. I have removed all MIN specific differences in optimisations levels compared to normal PicoMite

PicoMiteRP2040V6.03.00RC18.zip

@Peter,

a/ rc18 does not fix it, so your speed optimizations are sound.
b/ once the problem shows, powering from 5V does not fix it.
c/ when the problem is not showing, powering from 5V does not change that.

b/ and c/ there is no obvious relation between USB and the problem.

Regards,

Volhout

P.S. tested on a pico zero with WS2812, green is PIO FIFO ok, red PIO FIFO wrong.
Edited 2026-06-11 17:45 by Volhout

Hello Peter,

somewhere earlier I asked this, but it came up again during programming.
Is it possible in future versions to have possibility to jump forward/backward in a MP3/FLAC/WAV file which is currently playing?
For example with extra instructions like PLAY REWIND/FORWARD n (where n is in seconds).

G@bor

  Quote  

Is it possible in future versions to have possibility to jump forward/backward in a MP3/FLAC/WAV file which is currently playing?

No: the codec has to proceed linearly though the file

Peter,

I hadn't noticed this before.
Load a program with more than 100 lines.
Go to edit.
Go to line 100 or higher.
Use the arrow keys or mouse to position 9 and then to 10 or higher and look at the bottom info line. Fortunately, nothing shocking.
This only occurs with OPTION RESOLUTION 848.

Another point that may have been noted before, but I can't find it. If a program has been run in MODE 3, the mouse no longer works in EDIT. This is only the case for MODE 3; the others work normally.



Thanks for the nice and very useful additions to the current firmware.

Jan.

  Quote  

I hadn't noticed this before.
Load a program with more than 100 lines.
Go to edit.
Go to line 100 or higher.
Use the arrow keys or mouse to position 9 and then to 10 or higher and look at the bottom info line. Fortunately, nothing shocking.
This only occurs with OPTION RESOLUTION 848.

This behaviour is correct. The status line shows the most information that it can. When the combination of the line number and the column number increase that field enough to collide with the status line the next smallest version of the status line is then used.

  Quote  

Another point that may have been noted before, but I can't find it. If a program has been run in MODE 3, the mouse no longer works in EDIT. This is only the case for MODE 3; the others work normally.

This is a S/W limitation. In the editor, the mouse works by flipping the tile colours. There is no such mechanism available in mode 3, the character would have to be rewritten. The workaround is to switch back to mode 1 before editing.

Peter,

Thanks for the explanation. I understand.

Jan.

V6.03.00RC18

PicoMiteRP2040V6.03.00RC18.zip
PicoMiteRP2350V6.03.00RC18a.zip
PicoMiteRP2350V6.03.00RC18b.zip

New HDMIWEB build - Web, HDMI and USB host rolled into one
CPU SPEED command - change the cpu speed without a reboot
RESOLUTION command - change the HDMI resolution without a reboot
Various bugs fixed as identified previously in this thread

See the updated manual for full details
PicoMite_User_Manual.pdf

Use the HDMIWEB build with a touch screen 1024x600 HDMI monitor and the On screen keyboard, print a case and you have a fully functional self-contained MMBasic tablet computer

PicoMite MMBasic RP2040 V6.03.00RC18
Copyright 2011-2026 Geoff Graham
Copyright 2016-2026 Peter Mather

> option list
PicoMite MMBasic RP2040 V6.03.00RC18
OPTION SYSTEM I2C GP20,GP21
OPTION COLOURCODE ON
OPTION HEARTBEAT OFF
OPTION PICO OFF
OPTION CPUSPEED (KHz) 200000
OPTION DISPLAY 50, 100
OPTION F1 help
OPTION F5 list commands
OPTION F6 list functions
OPTION F7 list pins
OPTION F8 list system i2c
OPTION F9 fm
> list pins
GP0      1      OFF
GP1      2      OFF
GP2      4      OFF
GP3      5      OFF
Press any key to continue
GP4      6      OFF
GP5      7      OFF
GP6      9      OFF
Press any key to continue
GP7      10     OFF
GP8      11     OFF
GP9      12     OFF
Press any key to continue
> list pins
[LIBRARY] Dim _q%
Error : _Q already declared


Jan.

Also with RP2350.

Jan.
Edited 2026-06-13 08:17 by JanVolk

  JanVolk said  

Also with RP2350.

Jan.

I can't make either of my PicoCalcs do that with the Web 2350 version.