When tuning half step down # symbol does not appear

[masdeguzman-crypto]

When tuning half step down sharp symbol (#) not showing up.

[kosmo-sys]

Hey! Thanks for the feedback. Are you able to provide images so I can reproduce this?

[masdeguzman-crypto]

Thank you. Sorry had trouble uploading the picture via mobile. <img width="1196" height="880" alt="Image" src="https://github.com/user-attachments/assets/894f1d71-d895-426c-8f22-2ccc503b0d91" />

[rayng1]

I ran into this issue as well when patching my GP-50. The issue is because the script is a bit too aggressive with overwriting with 0x00 null byte padding. The string fields to be replaced are of variable length blocks, eg. 4, 8 or 12 bytes but the script assumes they are all null byte terminated. The variable $availableSpace is calculated using the distance to the next 0x00 null byte, so the loop keeps running and floods the rest of the block with zeros (0x00).

I asked AI which suggested a potential fix for the script. If I have time I'll provide a pull request but I won't have time to test it myself.

Workaround:
@masdeguzman-crypto
Worked for me with the below instructions for the GP-50 firmware but usual caveat of do at your own risk:
Using a hex editor (I used HxD - https://mh-nexus.de/en/hxd/) open up the English patched firmware .bin file for the GP-50. Change the first byte '00' at offset 00165FA0 with the value '23' (which represents '#'). See the screenshot where the byte to modify is highlighted in red. Once changed, save the file and update the firmware using this new file using Valeton Suite as per normal firmware update process. The sharp symbol should now show up.

[rayng1]

I wasn't able to raise a pull request so here's the suggest but completely untested code fix if anyone wants it:

        if ($engLen -le $availableSpace) {
        # UNTESTED: Wipe exactly the footprint of the original string block found by the scanner.
        # This completely cleans out old Chinese bytes without touching the next field.
        $maxOverwriteSpace = $current.length

        for ($j = 0; $j -lt $maxOverwriteSpace; $j++) {
            if ($j -lt $engLen) {
                $fileBytes[$current.offset + $j] = $engBytes[$j]
            } else {
                $fileBytes[$current.offset + $j] = 0x00
            }
        }
        $patched++

[kosmo-sys]

Thank you @rayng1 I'll test your work! Out of curiosity, why were you not able to pullrequest?

[rayng1]

Thanks for offering to test. I should add that I assumed that this issue is for the GP-50. The scripts for the other pedals work similarly but I don't know whether their firmware would share this same problem.

The Create Pull Request button is disabled for me. Might be a permissions issue.

[cornellpg]

New firmware version 1.1.0 released today.
Do you think this patch will need to be modified a lot?

[kosmo-sys]

I am not sure, but I don't think so. Let's hope they did not implement firmware hash verification, thou... are there any new features worth updating the firmware for?

[cornellpg]

New firmware updates Bluetooth midi and known issues.

The suite provides nam a2 conversion via snaptone and steem splitter (like moises app)

[kosmo-sys]

@cornellpg Try to use the patch as it is now from the branch gp-5-tested, you can force it to patch a firmware that is not of the same version. At worst you will see scrambled translations, at best the firmware has the same dictionary and will work the same. I have no time to test it right now, but please try it and let's us now!