The Dink Network

Heyo, with Robj's help I think we've beat Dink HD into shape, adding speed, bug and compatibility fixes. It should now be possible to fully enjoy Charlie's Legacy (and other dmods that push the limits) without issues.

If you want to try it: Latest beta version

Couple tips:

Tab = Extra speed. Shift+Tab = SUPER SPEED (EDIT: Oh, and ludicrous speed is Shift+Tab+Ctrl for some reason)

The debug menu (screenshot) is hidden, to get to it (it's not really for normal players) you need to Press F1 during gameplay, click "Debug tools" (upper right) and then "Toggle Debug DinkC". F1 to exit, then do the same thing again, but now you will see debug tools.

F4, F8 are hotkeys for quicksave, quickload.

I will roll it out to all the platforms Dink HD supports soon if it looks ok. Note: The about screen will say V1.97 Build 5. If you use it with Charlie's Legacy, be sure to download the latest version of that too.

Here's what was changed/fixed:

------ Change log for 1.97 ---------- (July 2022)

* BUGFIX: Fixed issue with black tiles at the start of Charlie's Legacy (Now properly rebuilds surfaces when dynamically switching from
8 bit to 32 bit color back buffer mode, it does this the first time it detects high color sprites in a dmod) (robj)
* BUGFIX: Fixed issue with pushing animation code that caused Charlie's Legacy to freeze dink on frame 1 when pushing (robj)
* Unless you're using touch controls, choosing an inventory item will no longer auto-close the menu, not really need for
keyboard/gamepad players (robj)
* load_script gives a warning to log if you try to load a script to a sprite that doesn't exist instead of crashing
* Added support for bmps using "bitfields" (to Proton), have never seen one in the wild, but Charlie's Legacy's map-03.bmp uses it
* BUGFIX: Unknown procedures give correct error in log, before it would include weird garbage
* HUGE speedup on script loading, fixes slowdowns in Charlie's Legacy sliding puzzles
(and pretty much any dmod that does intensive script processing)
* (Windows) I see possible issues with the script "return var" statement and stack problems, it doesn't properly
handle recursion well and will kill the stack if enough calls are made in the same loop. Instead of a real fix, I've increased the .exe's stack
from 1 mb to 50mb on Windows for now, will revisit issue later probably when handling mobile builds
* No longer writes "drawing hardmap" to the log
* BUGFIX: Fixed *.pdb open error which caused log.txt crash info to log less than it should have
* Changed every single sp_<something> command so it won't crash on a bad sprite index, will log it and return -1 to indicate an error
* BUGFIX: Fixed issue where reloading sequences in DMODS could eventually crash the game (robj)
* Audio assigned to an invalid sprite owner no longer causes crashes (robj)
* Say() command with invalid sprite no longer causes crashes (robj)
* A manual draw_status dinkc command now causes status bar variables to be updated instead of animating the differences, to match other dink engine versions (robj)
* BUGFIX: Reloading sprite sequences with longer ones no longer loses set_frame_frame info (robj)
* Added "Unfreeze Player" command to the secret debug menu

"If you use it with Charlie's Legacy, be sure to download the latest version of that too."

Just to add, Charlie's Legacy version 1.10 has been uploaded (couple of last minute fixes and stuff), might be a good idea wait until that version is available before testing it on the new DinkHD 1.97. It's just waiting on approval from Dan.

Edit: Charlie's Legacy version 1.10 is now approved and can be downloaded

Hey Seth, I have a bug to report. I was using my Martridge launcher to try to start Charlie's Legacy using DinkHD and ran into some issues regarding the -game argument. It seems this affects all versions of Dink HD.

Here are the problems:

1. If you use quotation marks when specifying the dmod path after -game, Dink HD adds a trailing / after the quotation marks and fails to launch the dmod. Example screenshot!

2. If you have a space in your path, even if you use quotation marks, the path will end at the first space! Example screenshot!

(In the second screenshot I just renamed the "DMODS" folder to "D M O D S" as an example.)

NOTE: If you don't use quotation marks, the path again will end with the first space, but in this case I suppose this is expected since the argument would clearly end there. FreeDink also fails to launch a dmod if the path contains a space and no quotation marks to delimit it.

EDIT: PS, these work fine in FreeDink

what do you have against f5 and f9 for quicksaving and loading? you were right there

DMOD path problems - thanks! I'll go fix it

Quicksave Hotkeys - Hmm, I agree, F5-F9 seems to be the most popular, and nobody at all uses F4... Ok, I'll move them, should decrease chances that it interferes with a global system hotkey players setup.

Ok, both those changes were added and the beta file (same URL as before) was updated. It will now say "Build 7" on the about screen.

hell yeah, we can save scum the bethesda approved way

>DMOD path problems - thanks! I'll go fix it

Nice, DMOD path problems are gone, thanks!

Edit: I had this one weird bug happen where the mouse cursor wouldn't show up in the DinkHD interface (when I press F1 for example). But I don't know why that happened and I can't seem to reproduce it now... just thought I'd mention it though.

Edit 2: PS, unrelated, but I really like the shadows in DinkHD.

Found another bug that might be annoying to players. Sometimes I've noticed when you press escape and load a save game(an actual one, not a save state), the "(Use F1 to bring up the Dink HD menu!)" text does not disappear - it stays on screen indefinitely. Once this happens, the only way to get it disappear is to restart the dmod.

It's strange, I can get it to happen rarely in most dmods. I have one Dmod(which isn't released on the DN - it's an updated version of push and pull I haven't released yet), where you can reproduce it every time, by saving with a cheat script, and then loading. So if you can't re-produce or figure it out, let me know, I'll send you the .dmod file and you can load it up and reproduce it 100% of the time by using the save game option in that choice menu.

>...Sometimes I've noticed when you press escape and load a save game(an actual one, not a save state), the "(Use F1 to bring up the Dink HD menu!)" text does not disappear - it stays on screen indefinitely.

Oh yeah! I noticed this once too, but I had no idea what I did to cause it and had no idea how to reproduce it.

Thanks, will check out the f1 message issue as well as the "losing mouse pointer" problem, would really like to fix that last one especially, happened to me several times, was hoping it was only an issue in debug mode but guess not!

Quick question. With the dawn of the Steam Deck, is there a chance to not only provide support for it but also add Dink Smallwood HD into the flathub repository so it can be installed through flatpak?

Flatpak being an universal package manager which runs the software in a sandbox environment with its own dependencies and permissions set by the developer. That way regardless of the linux distro or if you are on a Steam Deck, you can just simply install it and run.

I would imagine the issue would still be getting DinkHD to run on Linux in the first place... Unless you just WINE it?...

Hmm, I don't know anything about flatpak or steamdeck (don't have one yet). Maybe a better fit for freedink? I think personally I only want to support Win, Android, iOS and web/html5 right now.

Basically just keep them accessible and running well. If anyone else wanted to work on that I'd be happy to accept their work though. (the base framework does run on linux already via opengl or sdl, so it's more an issue with packaging builds, polish, and things like resolution/controller support)

Updated download to "Build 8" with the following fixes:

* BUGFIX: "Use F1 to bring up the Dink menu" message will no longer get stuck on the screen (Robj)
* BUGFIX: Fixed issue where sometimes the mouse cursor would stop showing up in the Dink HD F1 menu (ebilV)

Looks like we're just about there to make this a real release. Mobile releases should happen soon too.

Seth, this might seem like an odd request. Any chance for a DinkC command that disables(and re-enables) the use of save states/load states.

In certain situations, it will come as an advantage to the player not to have the option of using them... For instance, there's a place in Charlie's Legacy, where a certain fight was designed to be almost impossible, and if you die, the Dmod progresses anyway(after a Cutscene). During this fight I disabled the use of the escape menu so the player could not load. A player using DinkHD save states may never realise they are supposed to die because they keep state loading whenever they come close to death and just assume "damn this dmod is too hard, I'm quitting"

i kinda feel freedink should be able to natively run on the steam deck, but i haven't really looked at it much since its hella pricey

Ah, yeah, that happened to me actually, finally figured out I was supposed to die...

Ok, added. It works a lot like the "screenlock" command. (well, as in, you turn it on and off, not that it's locked to a particular screen. If you turn it on and don't turn it off, it will not allow state saves forever)

I tested it by override keys 1 and 2:

key-49.c:

void main(void)
{
show_quick_message("Save states disable for this section!");
set_disable_savestates(1);
}

key-50.c:

void main(void)
{
show_quick_message("Save states enabled again!");
set_disable_savestates(0);
}

You can pass -1 to not change it to just have it return the current value btw.

Note: I'm still allowing autosaves and "save on quit" to function, which operate separately than save states. (don't want anyone to lose their progress because suddenly save on quit doesn't work) It's not useful for save-scumming so should be the best of both worlds. The "save state disabled" state is remembered after continuing the game from a savestate in those cases. (confusing but I think it all makes sense)

If someone tries to save/load while it's disabled it will give a clear message that the dmod disabled it for this area.

Beta file updated: Build 9.

Here's what changed (ignore the android parts, won't matter here):

* NEW DINKC COMMAND: set_disable_savestates(1) to disabled. 0 to enable, -1 to return the current status. Note: Autosaves will continue to
function normally, and the current savestate status will be respected in autosaves as well
* NEW DINKC COMMAND: show_quick_message("Hello! `4This is red``. `2Green``."); shows a quick message over the dink screen, useful for showing quick messages, I used it
for testing stuff. Works with color codes.
* get_version() now returns 111. (dinkC version incremented by one due to added new scripting commands).
Reminder: you can use get_client_version if you want 197, the client exe version.
* If dinkc debug mode is on, now logs extra info about "external" calls being done
* Reworked dinkc to use much less stack memory, fixes issue on Android with nested "external" calls
* (Android) Can now use Xbox, PS4/PS5 controllers via bluetooth on Android, if they are paired, they just work. (note: turn down
the touch countrols GUI visible slider down in the settings if you don't like that cluttering the screen)

Hi Seth. Thanks for adding that.
2 things.

1. The about screen still says the script version is "110" not "111". Although the command does return 111, it's just the about page showing the wrong one.

2. I think the script loading speed up thing took a hit in this build somehow? Push and Pull and tile puzzle gone back to being laggy.

If it's to do with the Reworked dinkc to use much less stack memory, for android, and there's no other alternative, would it be possible to have it work as it did before if it's detected running on non-android platform? I dunno if that's possible.. I mean, even though it's laggy on android, it would still be complet-able(with patience.. the lag is real), but if the PC folks don't have to take that performance hit, since windows is capable of handling it as it was before, that'd be the best of both worlds... otherwise, it's back to the original issue of lagginess with the complex stuff, so I do hope this can be fixed

Thanks. Woah, speed hit? Argh, that's no good, I'll check it out and also fix the about screen.

Ok, please try the latest "Build 10" if you get a chance. It should be very close to the original speed, although maybe SLIGHTLY slower, if you can notice a difference, I'll just not use it in Windows builds like you mentioned.

BTW, I guess this is Dink HD specific, but I just remembered/realized in-line coloring works `4everywhere`` if you ever need that. Full list: https://growtopia.fandom.com/wiki/Chat#Text_Coloring

Besides the fixes, the only other thing tweaked in this build:

* Controller support improved on Windows and Android, 8Bit Do, Nintendo Pro, PS4,PS5,XBOX should be auto-mapped right. Start on controllers
now always opens the Dink HD menu, pushing left/right allows selecting save/load state easily. Select button will open
the classic Dink menu.

"Ok, please try the latest "Build 10" if you get a chance. It should be very close to the original speed, although maybe SLIGHTLY slower, if you can notice a difference, I'll just not use it in Windows builds like you mentioned."

It's much better, although still able to obviously feel the slowdown. This change will be much more play-able for phone/tablet at least now, but I vote to not use it for windows if that's an option (I have some other stuff in the works that I tested it on, and it is much better on the non-slowdowned version ) - if windows doesn't have to be held back by the phone/tablet android thing, then full speed ahead sir!

That in line text thing is cool. And, we can always run different DinkC lines based on get_version, so a dmod is compatible on all Dink versions, but still benefit from new features of DinkHD, if the player is using it. Awesome.

Ah, too bad. Ok, I'll keep the "stack hack" out of the Windows build.

I've updated the build to 11 - with that change. So this should be back at full speed. If it still seems slower it could be from a different change.

EDIT: Build 12 with a bit more gamepad/GUI work, this is the release candidate if all looks ok

holy shit, nice. i should reinstall dinkhd on mobile and try that out

"EDIT: Build 12 with a bit more gamepad/GUI work, this is the release candidate if all looks ok"

Perfect, and slowdown is gone again on windows.

Ok, I released V1.97 final version. ( https://www.rtsoft.com/dink/DinkSmallwoodHDInstaller.exe )

All Windows existing Dink HD's will prompt the user to update now.

Kind of dangerous, but I did some last minute fixes, one was to make the in-game DMOD browser allow more room to fit the text/description, also, forgot to mention another recent change: the default sorting for when you're choosing to download a DMOD is "Newest", so hopefully authors actively working on DMODS will get the most downloads.

Currently the #2 DMOD listed inside of Dink HD's dmod browser is "Version Checker" - I guess the "Development" category should probably stripped from the DMOD list that's being sent. Dan, you around? Although, I mean, it does work and let you play it so.. hrm.

Anyway, I guess that's a wrap, thanks! I expect the mobile versions to be updated with V1.97 soon. If any serious bugs are found I'll fix and bump it up to 1.98.

Looking forward to being able to test Charlie on the mobile release once it's live.

I have one request/suggestion regarding this V1.97 release.

Please upload the new version over here on TDN as well. https://www.dinknetwork.com/file/dink_smallwood_hd/

Last version here is 1.93 and I'm sure there are people still downloading it from here, would be a shame for them to get stuck with an older version because they don't know to look elsewhere.

Android version has been updated: https://play.google.com/store/apps/details?id=com.rtsoft.rtdink

Welp I got to the Push & Pull tutorial in Charlie's Legacy, found out how that turned out with the difference between it & windows version.
Tutorial Barrel pushing
I suspect this slowdown was as expected though, yet to reach the tile puzzles.

EDIT: We're finding the sound effects to be running at a slower speed/pitch than expected as well, and additionally when I use Fireball it plays sel1.wav for some weird reason. It's also playing barrel break when fireball hits anything; Robj suggests the sound files are playing at the wrong speeds (we screen-shared DinkHD over a Discord mobile call)

I think the slowdown for ExDeathevn is old tablet specs. I tried it on my android phone, slowdown wasn't too bad. There, but tolerable, as expected.

I did experience the sound speeds bug on my phone though. It makes the fireball sound like the select sound when you shoot.

I should add it's not all sounds being slowed down, or at least it seems that way.

Cool, very helpful. I'll take a closer look, I'm trying to get the iOS version ready, need to see if that has the same issue. iOS version doesn't seem to need the "stack hack" so at least it will be pushing barrels fast though.

Another minor bug? So during development, we removed the black backgrounds for inventory items & magic frames in the image files, but on the mobile status bar the black box appears behind the active magic again. Unsure if just mobile or not.

Good news though, the first Sliding Tile puzzle worked as intended, if a little slower on my devices (tablet & android phone).
I'll be continuing to note anything I see as I go, we initially were never expecting it to work on mobile so this makes for a good testing opportunity anyway.

"So during development, we removed the black backgrounds for inventory items & magic"

I'm gonna confirm this when I get home tonight, but I need to check if that actually took affect in the dmod at all on any dink engine. Looking at some previous screenshots the black backdrops of magic items are still there in windows DinkHD too. I'll test freedink tonight, I can't remember if they show up on there. If so, it's not an engine bug.

Edit: also checked some earlier screenshots taken from FreeDink, the black backdrops are indeed removed. I think DinkHD is automatically adding a black backdrop into the armed magic square on the status bar or something, behind the chosen image, because the backdrop isn't there for the icon when you open the inventory, it's only on the status bar.

It's not doing it for the weapon icons though, the black backdrops for those are gone on the stat bar as intended.

Double post for bump.

One difference I noticed in DinkHD. When you change &update_status to 0, it entirely blacks out the status bar, instead of keeping it displayed, but not updating(like in other Dink engines) until you change it back to 1 (which in DinkHD unblacks it out).

Not sure if fixing that is needed. But, just for the sake of consitency..

EDIT: I sent you a PM as well Seth. I'm after a list of all the added DinkC functions in Dink Smallwood HD, so I can update the online DinkC Reference. Cheers!

EDIT2: There's also a key-stroke inconsistency. In other engines, you can hold a key on the keyboard, and the key-##.c script will be constantly called while it's held down (even while using the normal controls, such as the arrow keys). Basically, you should be able to hold one key-## button, plus move around, punch, use magic, and it shouldn't cancel the calling of the script. I originally had a "hold z to pull" mechanic in push and pull in the earlier verisons, had to remove it because as soon as you press the arrow keys in DnkHD, it won't call the key-## script anymore.
I don't know if any other Dmods rely on being able to hold a custom keystroke and also move around, but DinkHD won't let you do it.
Holding the speedup key also cancels keystroke calls in DinkHD.

EDIT3: The alt key temporarily freezes the game - is this intended? Also the enter key does sometimes randomly force full screen for some reason.

Thanks!

Keys checking for a script only on keydown rather than constantly - I think this was by design for performance reasons? I think this feature originated in 1.08(?) by Dan so it was probably designed that way. I actually hope people don't rely on external keys because it doesn't work with gamepad controls, or touch controls on mobile. (Although if they have a bluetooth keyboard...) If it's vital my arm could be twisted to make it work like that, but changing might be bad if other dmods relied on it being run only on key-down. (holding it could run a script 10 times instead of once?) Random thought, the proper way might be to be calling key-##-release.c when they let go. (anything to get away from calling it every frame)

Alt freeze - yeah, I was about to say it's a windows thing (alt on windows shows the task bar settings in many apps - minimize/maximize/size) but then I tried it and noticed it doesn't show any menu initially (until you press down after hitting alt) so that's not right. I should fix..

Enter toggling fullscreen - Alt-Enter is kind of a windows standard to toggle fullscreen/windowed, Dink honors this. So probably you had hit Alt and enter together somehow? If enter by itself did it, that would be a bug.

Oh, before reading this I did release V1.98 for Win: https://www.rtsoft.com/dink/DinkSmallwoodHDInstaller.exe I didn't bother with beta since not too much was changed.

I've submitted 1.98 to iOS and Android as well. Mobile versions have a nifty new feature where if you use a gamepad the touch controls will hide themselves, so an iPhone + PS5 controller broadcasting with Airplay or whatever to a big screen tv is a pretty nice dink experience now.

------ Change log for 1.98 ---------- (August 2022)

* (iOS) Native Bluetooth controller & keyboard support added, same as Win and Android builds
* Added an emulated pointer where needed, all GUI screens can now be 100% controlled by controller or keyboard,
mouse/touch not required anymore
* DinkC addition: get_client_fork() added. Always returns 1.
* get_version() incremented one due to the above change. Now returns 112
* (iOS/Android) BUGFIX: show_bmp dinkc command will no longer fail on some dmods due to case sensitivity of linux based systems
* No longer drawing a black square behind the active magic icon. No idea why I was doing that. (Robc)

Just to clarify, DinkHD keys do constantly run on key being held down(actually more rapidly calling the key-## script than earlier engines)! My point was that pressing any other 'normal' control stops the held down external key from constantly calling in DinkHD, where as in other versions, normal controls don't do that(only another external key does), you can hold an external key while also moving around with arrows for example. So basically in other engines you could hold down a external key while moving/fighting/any of the normal controls, and it wouldn't interrupt the call. In DinkHD, external keys only works while Dink is stationary.
It's not a big deal though.

Ah, I see, weird. Pushing a key goes crazy and calls it over and over? Seems like that would cause a lot of problems and it would spawn 200 scripts if they were using wait()s and didn't return instantly? Not sure why a second key would stop the first one. I need to take a closer look, I suspect iOS/Android/Win all work differently, probably a good thing that feature isn't used too much.

Good news: 1.98 is now live on iOS/Android/Windows! Will keep noting reported issues, hopefully no showstoppers.

"Seems like that would cause a lot of problems and it would spawn 200 scripts if they were using wait()s and didn't return instantly?"

I haven't seen too many dmods that include waits in the key-## script, some include choice commands, but they tend to pause calling of the script until after the choice command goes away. I can't remember which Dmod, but I remember at least one that relied on this.. it's an old one.. I just can't remember it. But I do recall looking into the script and authors that use this tend to be script friendly (at least that one did), and increment a variable on the call of the script, and instantly "kill_this_task" if said variable is incremented (so further calls die instantly while it's active). I think authors should be responsible for knowing how certain calls work and making sure their dmod is compliant anyway. But yeh, I'd say there's only a couple of dmods at most that have used this in the past.

EDIT: I'm not getting a notification to update in DinkHD 1.97 BTW, and I'm pretty sure I'm using the official 1.97 build 1 release, not a beta, so it should be psetering me to update. Is 1.98 accidentally marked as a beta build or something? ..