This was driving me nuts.
When the number of modules in the folder is 1, the program gets stuck on load.
0 or more than 1 work fine.
The release version does this.

I played around with the directory-crawling code and made it not to crash but I have no idea exactly what was happening.

I have to send Hoffman my code so he can compare.

I've been using this monitor-less many times, it's very possible, with a few tweaks it would be just a perfect.

CIA interrupt is now running slightly slower under when called by the OS. Also CPU speed and fast ram look to be making an effect to this speed as well (runs different speeds on A1200 and A600 for example).

What we really want to achieve is solid timing at standard playback rate on all platforms, if possible.

@dwhinham wonder if this is something to do with the when the interrupt is confirmed?

Many pieces of Amiga software come with a nice hyperlinked AmigaGuide.

Perhaps the text of the PDF manual (#14) could be incorporated into an AmigaGuide, preferably in some kind of automated/scripted fashion to make it easy for us to keep the two in sync after updates.

Reported by nkrypth, apparently, an NTSC machine switched to 50Hz does not behave exactly as a proper PAL machine.

This needs testing, sounds real weird. I don't have an NTSC machine to test with and did not try with emulation.

We should come up with or pick a commonly-used code style guide to stick to throughout the project, which defines things such as (but not limited to):

• Tabs (and tab-stop width) vs spaces
  • Currently using tabs with tabstop of 4
• snake_case vs CamelCase
  • Currently using snake_case in the style of oldschool C
• Variable names
  • Descriptive, but not loo long
• Function names
  • To avoid naming conflicts with any external code we might bring in, I like the idea of projectname_modulename_functionname(), e.g. pt1210_file_open()
• Braces (first brace dropped-down vs first brace in-line)
  • Currently dropping down to a new line for the first brace to help clarify where blocks start/end
• Ensuring trailing whitespace is never present
  • Editors can be configured to auto-strip trailing whitespace

We can also employ tools such as clang-tidy to automatically detect when we stray from these and automatically fix them.

Editorconfig a nice tool which is built into many editors or available as an extension which reads an .editorconfig file and sets up things like indentation automatically.

the BPM of ProTracker is not 100% accurate, need to investigate why this is and if it is possible to make ti 100% accurate. if this can be achieved it can be added to a new config page.

Idea: Detect ARexx on OS2.x and above, and if the user enables this feature, provide an ARexx port to allow users to write scripts to automate PT-1210.

This could be used to interface PT-1210 with some kind of remote control (e.g. serial, network), stream track information/BPM, or other useful things.

It may also be useful for automated testing.

Add full Latin set characters, and icons for folder/file

I started this and maybe it's close or it to work.
I am using JOTD's CD32 pad code found in the JST source here: http://jotd.pagesperso-orange.fr/amiga.html

I included his code and I am just unable to wire it to the rest of the program. Hoffman will have to look at my version of the code when I send it to him.

I already devised a preliminary control scheme fr CD32 pad control. It involves shift keying:

L+R = shift keys
Pad Left: Nudge -
Pad Right: Nudge +
Pad Up: BPM +
Pad Down: BPM -
R+Pad Left: Nudge Hard -
R+Pad Right: Nudge Hard +
R+Pad Up: Pattern Fwd
R+Pad Down: Pattern Rwd
Red button: Loop FX on/off
Blue button: loop length cycle
Green button: slip mode
Yellow button: Set Loop In/Out/Off
L+Yellow button: Repitch on/off
Start button: Play/Pause
L+Start button: Cue
R+Start button: Set Cue

Of course this will need intensive testing and redoing.

I have a controller board to create a CD32 pad. I could even turn this into a bespoke controller!

It would be great if the original ASM source gets more comments.
I added what I discovered to my copy of the code, but I was mostly travelling blindly through it.

I promise to send Ian this code today so he can compare/merge.

It’d be neat to have each computer on a different colour, changing just some of the colours, like digits display colours and playhead colour.

Provide a configurable screensaver that can blank the screen after x seconds to prevent screen burn-in or prevent distractions.

MFWA (Make F00 Work Again)

Reported by Pekka Korolainen:
"Play a sample on a channel, 3F0 into another sample and the waveform displayed is from the previous sample (minor and also not annoying)"

I thought we had an issue here for this, as we discussed it, if we do, please close this one.

Have the mutes shut audio off immediately instead of not letting further samples play while muted. I guess this would also be tied to my "gater fx" request.

Add a feature to the file browser to show which songs have already have been played in this session, a bit like Traktor.

This could be as simple as a small icon next to played songs.

The indicators should survive when the user navigates up and down the file system hierarchy.

Idea for implementation:

• Implement a fast hashing function (e.g. Pearson hashing) that can scan a module header and create a unique hash code based on its file name, file size, song title, and sample names (n.b. whole module header).
• For every song that has been played for more than $SOME_AMOUNT_OF patterns, add its hash into a global in-memory "played" list. This could be a fixed length, or dynamically-growing.
• When refreshing the file browser, hash each song being scanned. If the resulting hash code is in the "played" list, draw its "played" icon in the file browser.

We could do this by shift-pressing the already present channel mutes. Shift-1 will solo 1, etc.
Maybe another shortcut to un-mute all muted channels at once.

On load, the cue marker (red bit on the sequence band at the beginning) shows up, but after you set it anywhere else, it doesn't show up anymore. Everything works, it's just the visual aid disappeared.

Take all the text I already made for the PDF manual. revise it to match current features, and add it as TXT file.

these have happened in the wild but have not been able to be captured or reproduced as yet. need to consider how to build a hardy debug version which we can push out for testing which will capture the exceptions and give us some idea of what is going on.

convert the majority of code to C enabling easily development of all the new features we want to put in. in the process we need to start stripping out all the nasty system killy stuff.

Update the splash screen for v1.1 (I think I almost have it already) and include it as a separate iff file to be displayed on boot with some kinda util on the ADF version.

Make the currently active frequency table locked and turn off repitch at that point (for harmonic mixing)

When you boot straight to PT-1210 and you remove the splash screen (like I had or test purposes), in OS2.x+, the program will not initialize the screen properly because the display initialization code seems to be in the display code of the splash screen.

This never happens on OS 1.x because it will always open a shell screen when booting off a startup-sequence. "Fix"was for me to output to shell so it would open the screen before booting the program (with like an echo comand).

This will probably get squashed with the refactoring but I thought I'd mention it in case anyone lands on it.

Allow the user to configure PT-1210 with an .INI-format configuration file that's human-readable and easy to edit with standard Amiga or PC text editors.

Idea: could these options be mirrored in icon ToolTypes, too? Perhaps ToolTypes would take priority, followed by config file, followed by default values.

Initial list of options to implement (checked = implemented):

general

• splash: Enable splash screen on startup (on/off, default: on)
• close_wb: Attempt to close Workbench when PT-1210 starts (on/off, default: off)

dir_cache

• mode: Set directory caching mode: 0 = disabled, 1 = auto, 2 = manual (0/1/2, default: 1)
• whitelist: Comma-separated list of paths to include when caching (manual mode only) (default: empty)
• blacklist: Comma-separated list of paths to exclude when caching (manual and automatic modes) (default: Ram Disk:*)

player

• soft_nudge_bpm: BPM to add/subtract when using the 'soft' nudge function (default: 1)
• hard_nudge_bpm: BPM to add/subtract when using the 'hard' nudge function (default: 6)

keyboard

• repeat_delay: Amount of time, in milliseconds, that a key must be held before it repeats (default: 500)
• repeat_rate: Period, in milliseconds, at which a held key repeats (default: 1000)

gameport

• port_0_mode: Mouse port mode: 0 = disabled, 1 = auto detect, 2 = force joystick, 3 = force CD32 pad. (0/1/2/3, default: 0)
• port_1_mode: Joystick port mode: 0 = disabled, 1 = auto detect, 2 = force joystick, 3 = force CD32 pad. (0/1/2/3, default: 1)

bindings

• A list of keyboard/joystick action bindings in some reasonable format.
  Example: LEFT_SHIFT + UP = pitch_up_fine
  This section should contain the entire default binding list, maybe commented out so the user can easily tweak as they please. PT-1210 will contain the default bindings internally in case that a config file is not found. If this section contains at least one valid binding, the internal defaults should be ignored, and it's then on the user to map all the keys they want in this file.

New line loop code to monitor sample positions, means looped drum samples will re-trigger when looped half way through.

I have this 90% written, just need formatting and graphics.

we have a function where you can write !FRMxx in sample name 31 to set the number of frames per best. this has to be done manually at the moment in protracker. would be great if within pt-1210 you could pull up a mod config page and set this inside the software and ultimately save that config inside the module.

We now open the PT-1210 screen in a semi-system-friendly fashion by bringing up an Intuition screen, but then hijacking its copperlists every VBlank.

By rewriting the copper lists to be system-friendly, the ability to flip between PT-1210 and Workbench will "just work", and the screen can be pulled down to reveal Workbench for true friendly multitasking.

I cannot consistently reproduce this one but I can 100% have it happen every time I have a session :/
Sometimes module loading crashes the software with a guru. It must be some kind of memory allocation issue?

Currently we load the entire mod into chip ram which isn't great if the user has a stack of fast ram.

would require..

1. load initial header - discover how many patterns there are
2. load pattern data
3. load remaining data into chip
4. modify PT player init to take all three pointers.

loading from floppy on kickstart 1.2 / 1.3 does not work once the system is up and running. before the system is taken down however, it does work fine (loads the list of modules on start-up).

suspect this is because kickstart 1.2 / 1.3 use the blitter to decode the MFM floppy disk buffer but we are hammering the blitter in an unfriendly way, don't think we even own or disown the blitter from the OS.

once all interrupts are system friendly and the UI is being updated in a friendly manner, then this "should" resolve itself.

would be nice.. but also need to fix the scopes and some other stuff as they are hardwired to 50hz.

I liked the idea of an "About" screen/page, could we still do it but text only? That way it would not take much RAM:

I just had a crazy super low priority idea, hear me out.

I heard it's pretty easy to control an I2C LED display using the serial port.
Could we have it display the tune title we're playing or something?

It would work great for a headless design, with a little LCD showing some info, or even connecting it to a "ticker"-like large LCD that shows the damn tune I am playing, so people on YouTube stop asking me for a tracklist :P

I guess even just sending the data through serial as a string could be handy if you hook it to a PC and make a script that inserts it into OBS or whatever...

Anyway SUPER low priority this!

Idea: Under WB 2.x+, enable an AppIcon on the Workbench.

This could be used for say, if the user wants to use Workbench or some other graphical file management to locate their modules. Modules could then be dragged-and-dropped onto the PT-1210 icon to load them into the deck.

This feature would require #45 to be completed first.

Reported by Pekka Korolainen:

"If I play a tune with an EFx, x != 0; and load another mod and there was no EF0 in the previous mod, the EFx carries over (although EFx is the most awesomest effect, this sounds pretty bad sometimes)."

Reported by Barnabas in Discord, full details here:
https://forum.icomp.de/index.php?thread/1393-indivision-ecs-v2-with-pt1210-software/&postID=12528#post12528

Can Ctrl be used as a modifier? If so, it'd be nice to be able to move +/- 10 patterns if we do Ctrl plus the +/- keys. Very long mods can be traveled thru faster that way.

add function to keep track of the sample pointers on each line as the track plays, then when line looping is activated, it can re-trigger drum loops at the correct point if no note exists on that line.

note: work out formula to take into account period and bpm to determine how far we've moved since last frame. similar to the the way the scopes work.

config options for a module

• frames per beat ( default is 24 )
• bpm ( detected and stored but can be edited )
• track name edit
• artist name edit ( detect by finding first sample name with # )
• cue points ( yes I know they don't exist yet )

all the info we'll store in the sample names so writing the data back to the module should means only writing the header (should be dead quick). need to decide on a flag in the header which denotes all data should exist because it's been saved by PT-1210

side note, possibly option to knock up super quick C# app for doing this function in windows.

Once the config file is in place (#10) it would be good to be able to set these parameters independently of the system settings.

Pitch range selection like in a CDJ (+/- 8%, 10%, 16%, Wide (20-300 BPM))

Currently PT-1210 only scans the current folder and displays only modules within that folder. Needs a proper file / folder browser including listing drives

i.e. a quick stop, not immediate audio cut off, say 1 second to slow BPM to zero and stop track

Very small dirs work OK but longer ones get stuck in "reading directory".
Sometimes after a while waiting like that, screen gets corrupted, some "software failure" requester can be made out, and then Amiga reboots with guru.

Test iso provided in another issue.

Cuts audio at a predetermined rate i.e. every other line to get a cross fader/stutter type effect.
Should be maybe another mode of the line looper? Mute volume every X step?

See if I can make it work with the space we already have or if I need one extra block of pixels underneath it, eating up the tracker display.

create brand new UI data format so graphics engine does not work from hard blits and multiple incbins. instead create a theme template and graphic tool to allow for inclusion of any theme.

currently the graphics are split into lots of tiny files and the blit functions in asm expect them to be in a particular place and a partiuclar size. also there are a lot of copper tricks on the HUD to try and save memory which we probably don't need. removing these will un-complicate the screen setup we have and also allow for easy palette configs to be loaded or even included with the theme.