Does this do it?


PicoMiteRP2040VGAV6.00.02RC19.zip

Hi Peter,

Unfortunately not
after ctrl-c
print mm.info(tile height) gives 16 (the value set by logic analyzer). Not 12 (matching the font).

Volhout

IGNORE
Edited 2025-05-03 02:36 by matherp

Harm

I think this does what you want but there is an interesting issue which is certainly there back to V6.00.01 and probably before.

PicoMiteRP2040VGAV6.00.02RC19.zip

Run LA_24_2 and you should find all is OK on exit with the tile height back to 12

Now set a different default foreground and background colour - it is the background being set that causes the issue

OPTION LCDPANEL CONSOLE 1,rgb(red),rgb(blue)

You will see the program still runs fine but the block colours in some areas are wrong. It would be really helpful if you could diagnose why. Is it some tiles that you are not initialising after the height change? Or can you identify what call you are making that the firmware isn't doing what we expect.
Previous versions wouldn't have restored the foreground and background colours properly like this one but the effect on your program is exactly the same

Thanks
Edited 2025-05-03 03:24 by matherp

@Peter,

This version works.
I looked at the logic analyzer, and in line 437 I set the height to 16, and in 438 I initialize all tiles to white/black. Later some tiles are changed to yellow, green, and red to support the graphics.



But it is using a strange combination of tiles, graphical elements (banner2.bmp) and framebuffer. Earlier versions of logic analyzer used banner.bmp, which was an inverted version of banner2.bmp. Somewhere in the 5.07 -> 5.09 I needed to invert the banner to achieve the same visible appearance. This may partly be caused by Martin.H's original artwork for the logic analyzer that used swapped foreground and background.

"Sub set screen" is executed AFTER the buffers (especially the ring buffer) are initialized, so the frame buffer may be located on an uncommon address.

At the moment I have no idea what the impact on the background is. I replaced the colour statement (above) with different colors, but all keeps working alright.

I made the option change you hinted, and can see the same corruption in colors.

Volhout
Edited 2025-05-03 05:20 by Volhout

@Peter,

I think I have an idea. I do not use framebuffer L. I use framebuffer F. But it looks like there is something when the background color is not 0 (0 is default transparent for L). Only where text is written with defined for/background color you get the correct colors.

Am I making sense ?

Volhout

Peter,

Maybe do not change anything anymore. It wprks now, there is a known anomaly, that no one ever encountered before, and it only impacts tile based color vga.

I suggest to not break potentiallu critical vga timing by fixing an issue no one will ever encounter.

Volhout

IGNORE
Edited 2025-05-03 19:34 by matherp

Harm

I've found the issue. You are issuing the TILE HEIGHT command after the FRAMEBUFFER WRITE command. It will break your program, albeit trivially, but I'm going to make using TILE HEIGHT when FRAMEBUFFER WRITE is set an error to avoid these sort of unexpected issues.

  homa said  

I'm back from my short holiday. And I'm looking forward to testing again. If there are still so many great additions, will the textbuffer layer be added after all?  

Hi Peter, have I missed an answer? I would be delighted.

  matherp said  

Harm

I've found the issue. You are issuing the TILE HEIGHT command after the FRAMEBUFFER WRITE command. It will break your program, albeit trivially, but I'm going to make using TILE HEIGHT when FRAMEBUFFER WRITE is set an error to avoid these sort of unexpected issues.

Wauw, I never would have thought that to be an issue. The frame buffers are both single bit, and the tiles are mixed in in the second arm. But I do acknowledge that changing the tole height reconfigures the memory, since there are fewer tiles.

Thanks for finding it.

Volhoit

V6.00.02RC20 is available on

https://geoffg.net/Downloads/picomite/PicoMite_Beta.zip

Pulls together a number of fixes in the various RC19 downloads

Fixes about a gazillion bugs when using a default background colour other than black in VGA and HDMI versions particularly in mode 1

Trying to change the Tile height when using FRAMEBUFFER WRITE now gives an error. Previously it could cause undefined and difficult to diagnose corruption of the intended display

Enables OPTION DEFAULT COLOURS fc, bc for all versions
fc and bc can be one of :        
WHITE,YELLOW,LILAC,BROWN,FUCHSIA,
RUST,MAGENTA,RED,CYAN,GREEN,CERULEAN,
MIDGREEN,COBALT,MYRTLE,BLUE,BLACK

Note no RGB() and no quotes

Changes OPTION LIST to show the CPU speed appended to the resolution for VGA and HDMI versions

HELP

I have no more HDMI/DVI output after startup! Reset does not help either! Only an >edit leads to an image output. After leaving the editor, I also have a prompt >



and the mistakes continue:

Edited 2025-05-04 05:01 by homa

Try clearing flash - I guarantee it works properly if loaded onto a cleaned Pico

Note to everyone - clear flash before loading this version
Edited 2025-05-04 05:21 by matherp

No, unfortunately not.
I just nuked the Pico with this: "Clear_flash.uf2"

First a screen display came up.
The same two errors after: OPTION RESET HDMIUSB



Addendum: The display works now ?!? But the error with the mp3 playback remains :-(

Matthias
Edited 2025-05-04 06:17 by homa

The mp3 error means a corrupted mp3 file. Works fine on my HDMIUSB.

Sorry, but these are the same mp3 files as in the previous release versions. Until RC-18, exactly these mp3 files were played back (b:/sdcard)... so far none of them work. The MOD files work without any problems.

I can install RC-18 and give it a try ...

ok I only have the RC-18 and then the RC-15 on my PC! In the RC-18 I also have this mp3 error, in the RC-15 the file is played back from the SD card without any problems.


Edited 2025-05-04 08:11 by homa

Hi Peter,
RE:RC20
Does the
#define USEBACKSPACE
that you temporarily took out resolving the previous issue need to go back in. That alternate code was from way back when I was developing this and does not work properly on the console when editing a wrapped command or if MM.PROMPT is in use with a different length for the prompt.




Also Draw.c void DisplayPutC(char c)  the red line needs to be commented out

Edited 2025-05-04 12:23 by disco4now

homa

I updated the mp3 codec to the latest version somewhere between RC15 and RC20. Both versions work fine for me. Is there an mp3 file you can post so I can try and identify the issue?

Gerry

Will do.

UPDATE

This version reverts to the older codec and includes Gerry's change - please report


PicoMiteHDMIUSBV6.00.02RC20.zip
Edited 2025-05-04 17:33 by matherp

Peter,

RC20 is fine for 2040 VGA PS2 platform.
Tested the majority of the programs.

Volhout