_PKL_main.ahk

;;  ================================================================================================================================================
;;  EPiKaL PKL - EPKL
;;  Portable Keyboard Layout (Máté Farkas, -2010)   [https://github.com/Portable-Keyboard-Layout]
;;  edition DreymaR    (Øystein Bech-Aase, 2015-)   [https://github.com/DreymaR/BigBagKbdTrixPKL]
;;  ================================================================================================================================================
;

/*
;;  ########################   TODO  »-->   ########################

WIPs: Recent and/or ongoing issues and projects
2FIX: Unresolved issues and bugs
NEXT: Improvements I'd really like to get around to soon
TODO: Improvements that are deemed worthy and useful, sometime
HOLD: Thoughts and suggestions that weren't that good after all, or currently infeasible

;;  ================================================================================================================================================
;;  eD WIPs/2FIX:

WIPs: Send fn() antics study. Can we make a SendInput call separate of the Key event?
		- https://discord.com/channels/115993023636176902/653362249687105536/1326675189353943050
		- Second example:
			b := Buffer(16 + A_PtrSize * 3, 0)
			NumPut(
			  "UInt" ,  1,      ; type = INPUT_KEYBOARD
			  "UInt" ,  0,      ; padding
			  "Short",  0,      ; wVK
			  "Short",  0x03B1, ; wSC
			  "UInt" ,  4,      ; dwFlags = KEYEVENTF_UNICODE
			  "UInt" ,  0,      ; time
			  "UPtr" ,  0,      ; dwExtraInfo = NULL
			  b
			)
			DllCall("SendInput", "UInt", 1, "Ptr", b, "Int", 16 + A_PtrSize*3)
		- https://discord.com/channels/115993023636176902/653362249687105536/1327235887553314838
		- Example:
			MySendInput() {
				INPUT_KEYBOARD  := 1
				KEYEVENTF_KEYUP := 0x02
				cbSize          := 40
				cInputs         := 80
				inputs          := Buffer(cbSize * cInputs, 0)
				
				loop cInputs {
					NumPut(
						"UInt"  ,   INPUT_KEYBOARD, ; type
						"UInt"  ,   0,              ; padd
						"UShort",   0x41,           ; wVk (VK_A)
						"UShort",   0,              ; wScan
						"UInt"  ,   Mod(A_Index, 2) == 0 ? KEYEVENTF_KEYUP : 0, ; dwFlags
						inputs,
						(A_Index-1) * cbSize) ; offset
				}
				
				DllCall("SendInput", "UInt", cInputs, "Ptr", inputs, "Int", cbSize)
			}

NEXT: Allow a BaseLayout stack: Variant,Options/Script,Base... ?
		- Make BaseVariants so we don't have to repeat ourselves for locales. The Layout.ini could just hold the ergo remaps.
		- The Cmk-Bul and Cmk-Ukr BaseVariants could for instance base itself on the Cmk-Kyr BaseLayout.
		- Just the Variant level, for now? Or could two more levels be nice? For instance, one locale plus one with, e.g., extra composes?
		- If going with multi-level, guard against infinite recursion (if already in stack, Continue). Limit LayStack depth to a few more layers?
		- Figure out a way to sort out the img_ entries too, without manually editing all of them? Soft/hard? Extend(@X)/Geometric(@H)?

WIPs: Galliard/Gralmak ortho, w/o AngleWide.

2FIX: The Shift key is often lost now, forcing a refresh? Only for Ext-Shift?
		- Could it be because some key combos change system layout now? (How?)
		- I also keep losing Ext Ctrl+Tab? What gives?

2FIX: PowerString name capitalization?
		- Composing, say, `say'pkl` and `say'Pkl` both output the ¶saypkl PowerString. Case should matter.
		- Checked whether a ¶sayPkl entry on the former sequence would work; it doesn't?
		- This may be another case of needing to make PklIniRead generally case sensitive, as with DKs?!
		- Only Composer respects case now, since it uses pklIniSect()?
		- For now, circumvented the problem by renaming the capitalized PowerString like `¶say-Pkl`.

WIPs: Detect OS VK codes for all keys instead of just a select subset, so OS layouts like AZERTY and Colemak-CAWS work as they should.

WIPs: With SC remaps, can we now actually remap the System layout? For instance, passthrough the OS layout but add AngleWideSym to it?!?
		- No, doesn't work; the SC don't get remapped at all. Ah well.
		- Consider which System mods to support. It may not make sense to add Curl there? But I want the right Extend etc.

WIPs: The CoDeKey sends repeated spaces when held down. Is this desirable? Could we specify no output by default for a DK?
WIPs: Ensure PrtScn is sent right for the CoDeKey and other DKs. Need PrtScn (all active windows), Alt+PrtScn (active window) and Win+PrtScn (full screen)
WIPs: Check out https://www.autohotkey.com/boards/viewtopic.php?f=6&t=77668&sid=15853dc42db4a0cc45ec7f6ce059c2dc about image flicker.
		- May not work with WinSet, Transparent; I'm using that with the Help Images.

WIPs: Add "What about gaming?" to README. Explain send method vs VK (also Compose etc). Mention MSKLC CAWS and SharpKeys.
WIPs: Introduce the marvelous Compose key in the README! Need more documentation on its merits. Also the new CoDeKey (dual-role Compose/Dead Key).
		- Become a Great Composer!

WIPs: Instead of doing the atKbdType() this-and-that routine, make a fn to interpret all @ codes and add it as a switch for pklIniRead()?
		- This would allow the use of all @ codes in all LayStack files

WIPs: "Add Layout" functionality in GUI, to select multiple active layouts without editing files manually.
		- Use the ComboBox functionality, that lets you have a DDL with a manually editable field on top.
		- Use an Add button? The button adds layout, line becomes <lay1>, then add is grayed out until something's changed. Could I avoid an extra button?
		- Or... a cheeky Join button that uses RegExReplace to merge the topmost two GUI override entries?! Too risky and error-prone for newbs.

WIPs: In the Janitor timer: Update the OS dead keys and OEM VKs as necessary. Register current LID and check for changes.

WIPs: Revisit the ISO key for several locale variants as the new Compose key is so powerful. Spanish? Probably not Scandi/German? Or?

WIPs: Make README.md for the main layout and layout variant folders, so they may be showcased on the GitHub site.
		- This way, people may read, e.g., IndyRad/QI analysis on the GitHub page in Markdown rather than the unattractive comment-in-file format.
		- Update correspondence between the Locale Forum topic and these pages: Link to EPKL in the topic, get info from the topic.

WIPs: Mother-of-DKs (MoDK), e.g., on Extend tap! Near endless possibilities, especially if dead keys can chain.
		- MoDK idea: Tap Ext for chaining DK layer (e.g., {Ext,a,e} for e acute – é?). But how best to organize them? Mnemonically is not so ergonomic.

WIPs: Dual-role modifiers. Allow home row modifiers like for instance Dusty from Discord uses: ARST and OIEN can be Alt/Win/Shift/Ctrl when held. Define both KeyDn/Up.
		- In EPKL_Settings, set a tapOrModTapTime. In layout, use SC### = VK/ModName first entries. The key works normally when tapped, and the Mod is stored separately.
		- Redefine the dual-role Extend key as a generic tapOrMod key. Treating Extend fully as a mod, it can also be ToM (or sticky?).
		- 2FIX: ToM-tap gets transposed when typing fast, the key is sluggish. But if the tap time is set too low, the key can't be tapped instead.
			- To fix this, registered interruption. So if something is hit before the mod timer the ToM tap is handled immediately.
			- However, Spc isn't handled correctly!? It still gets transposed.
		- Make a stack of active ToM keys? Ensuring that they get popped correctly. Nah...?
		- Should I support multi-ToM or not? Maybe two, but would need another timer then like with OSM.

;;  ================================================================================================================================================
;;  eD 2FIX:

2FIX: Repeat/Compose don't work on DK output?
		- Example: Typing {CoDeKey, [, Repeat} outputs`å[`.
		- Repeat works after composes. Not from OS DKs: Repeats the last non-DK press w/ Cmk-eD2VK. `ää` `áá` as it should w/ Cmk-eD.
		- Could be solved by adding DK output to the Compose queue. Might cause some trouble w/ how Back handles the queue then, but that's minor?

2FIX: SwiSh/FliCK modifiers don't stay active while held but effectivly become one-shot. And AltGr messes w/ them. Happened both on QW_LG and QWRCT.
		- The vmods don't need to be sticky for this to happen.
		- Are they turned off somewhere on release? That'd account for them working only once.

2FIX: Composes with apostrophe not working with eD2VK?!?
		- `^a` produces â but `'e` not é, etc. The culprit is the ' not being accessible from the underlying layout (registers as `o`).

2FIX: Holding MoDK-Ext then releasing it activates @ext0 even though Ext was held for > 1 s (over tapModTime = 200 ms).
		- If the Ext-hold is used w/ another key, it goes back as it should.
		- It'd be nice to make an Extend key KeyUp deactivate any MoDK layers. But how to deactivate DK layers like that?

2FIX: The `Ctrl+Shift+3` shortcut doesn't always suspend EPKL.
		- Seems it works after the first suspension.
		- Try to suspend/unsuspend EPKL once at startup then? Is that possible?
		- But WHY would this be the case?

2FIX: Releasing an Ext# layer leaves it active for a ToM timer duration.
		- With a dual-function Ext key, activate first Ext2 then quickly Ext1. Ext2 will stay active for one ToM timer.
		- https://github.com/DreymaR/BigBagKbdTrixPKL/issues/65
		- When I tried setting a high ToM Timer duration to test it, I couldn't activate Ext2 before the timer expired.
			- This is because of the tap (MoDK) getting registered, understandably. Would a working ToM interrupt help?
			- It is as it should be, I guess?

2FIX: DK detection works for my eD2VK layout, but not its pure VK counterpart. What gives?!
		- Is the BaseLayout wrong somehow? No, copying a key mapping from the Colemak-VK BaseLayout to the eD(2VK) Layout_Override preserves DK functionality.
		- Also, do DK detection when it's set to auto, and the old way otherwise? Then, what about when there is a table entry?
		- Call getWinLayDKs() (in pkl_deadkey) more robustly? No, it's the test for "śķιᶈForDK" in pkl_keypress that matters.
			- But keyPressed( HKey ) in pkl_keypress handles eD2VK and VK keys the same? So why then is the DK ruined by one and not the other?

2FIX: For the NNO WinLay, it registers SC00D as "1" and SC01B as "0:6"; they should be "1:6" (àá) and "0:1:6" (äâã), resp.?! How come some states get lost?!
		- Both are missing their state 6 dead key, acute and tilde respectively. There are no other dead keys in the layout. The Cmk-eD layout gets state 6 registered.
		- Might using ToUnicodeEx make a difference?
		- Reverting to listing DKs in the settings sounds like a defeat now...
		- Test it on the Colemak-eD MSKLC, since it has a ton of DKs? Only on levels 0:1 or 6 though.

2FIX: pkl_init runs through the layout twice. Is that really necessary, or does it simply double startup time?! Why does it even do this, again?!

2FIX: A ToM Caps/Ext key stops working as CapsLock after a while.
		- Initially, both work fine but after a while only the Extend-on-hold functions as it should.
		- Ext+Esc still toggles CapsLock as expected, and Caps-tap will turn that off. But not on, once it stops working.
		- https://www.reddit.com/r/Colemak/comments/14tmlvj/how_do_i_change_epkls_ext_key_to_say_lshcaps/
		- Could the mod-up sent on Extend key release be involved?

2FIX: When holding Extend-mousing for long with Timerless EPKL, there is still a hotkey queue. Probably the AHK hotkey buffer itself.
		- Problem: Once the queue is full, normal keypresses/letters start to occur. Occurs after ~2 s of Extend-mousing holding down the keys.
		- Is there a way of purging the actual AHK hotkey buffer? Or could changing its settings help?

2FIX: Somehow, the MSKLC Colemak[eD] does ð but not Đ? Others are okay it appears. Affects key mapped (eD2VK, System…) layouts. All other mappings seem okay.

2FIX: System mapping the QWP_# keys makes them ignore NumLock state?! Not sure how that works, but it's a tricky issue when one SC caters to two VK codes.

2FIX: The findWinLayVKs() fn is doing something wrong now? Trying to use the whole SCVKdic produces lots of strange entries...?!
		- Maybe I'm thinking all wrong about this though! There are two different issues at play: Where the OEM keys actually are, and how to remap keys.
		- Therefore, OEM keys should probably be treated differently from remapped keys (AZERTY, Cmk-CAWS etc). In some cases, a key can be both! Char-to-VK?
2FIX: SC remap the OEMdic, or layouts with ergo remaps will get it wrong. Example: Ctrl+Z on Angle stopped working when remapping QW_LG VK by SC.
		- In pkl_init, make a pdic[SC] = VK where SC is the remapped SC codes for the OEM keys, and VK what VK they're mapped to (or -1 if VKey mapped)
		- And/or a VK(ANSI)-to-VK(OS-layout) remap pdic?
		- Just detect every single VK code from the OS layout: It'd fix all our VK troubles, and account for such things as my CAWS OS layout.

2FIX: Hiding a DK image triggered by an AltGr+<key> DK fails: The AltGr help image gets stuck instead if it happens too fast. Affects hiding 'DKs'.

2FIX: Some new DK sequences don't work, like `~22A2   =  ~22AC	; ⊢ ⇒ ⊬` {DK_/,DK_=,g}. Others like `~2228   =  ~22BD	; ∨ ⇒ ⊽` {DK_/,DK_=,v} work. What gives?
		- Also iota/upsilon with dialytika and tonos don't work...?

2FIX: When selecting downwards with Extend and then using Extend-copy, sometimes an 'EXT' character (?) is made instead.

2FIX: Win+V can't paste when using ergo-modded layouts like AWide. However, with CAWS and Vanilla it works.
		- Is this because of the VK detection making an error? The ones that work both have V in its old place.

TEST: To avoid DK images stuck in the AltGr state, use a slight delay before showing the image if it's DK? It's a dirty hack, but could it help?
		- Would destroying the GUI on DK activation help at all?

2FIX: Every now and then (while using Extend?) EPKL becomes unresponsive to hotkeys and, e.g., changing tabs. Sometimes needs a menu Refresh/Restart.
		- No good ideas what causes this! It's annoying and happens too often.
2FIX: There are many composes with apostrophe; these may cause trouble for the CoDeKey when typing, e.g., `pow'r`. Move all acutes to, e.g., `''r`?
2FIX: Help images show 3–4× at startup with a slightly longer Sleep to hopefully avoid a minimize-to-taskbar bug on the first hide image.
		- It still doesn't work as it should, but the problem is hard to reproduce.
2FIX: Looks like there are multiple EPKL instances in the Tray now? Is that true? Can it be GUI windows? Can it be refresh related? Mouseover removes them.
2FIX: Ext-Shift may get stuck until Ext is released. Not sure exactly how.
2FIX: Help images for Colemak-Mirror don't show the apostrophe on AltGr even though it's functional and defined equivalently to the base state one.
		- Debug on 6_BS doesn't show any differences; looks like &quot; is still generated.
FIXED: Removed pressing LCtrl for AltGr (as in pkl_keypress.ahk now!). And changed to {Text} send.
		- Does it fix the problem with upgrading to a newer AHK version?!? No! LCtrl still gets stuck upon AltGr in AHK v1.1.28+.
2FIX: Setting a hotkey to, e.g., <^<+6 (LeftCtrl & LeftShift & 6) doesn't work.
2FIX: If a DK is selected very fast, the AltGr DK state image may get stuck until release. This happened after adding the DK img refresh-once timer?
		- Renamed any state6 DK images that contained only a base key release on Spc, to miminize this issue. DKs like Ogonek still have it.
2FIX: The ToM MoDK Ext doesn't always take when tapped quickly. Say I have period on {Ext-tap,i}. I'll sometimes get i and/or a space instead.
		- Seems that {tap-Ext,i} very fast doesn't take (producing i or nothing instead of ing)? Unrelated to the ToM term.
2FIX: Mapping a key to a modifier makes it one-shot?!
2FIX: Redo the AltGr implementation.
		- Make a mapping for LCtrl & RAlt, with the layout alias AltGr?! That'd pick up the OS AltGr, and we can then do what we like with it.
		- Treat EPKL AltGr as a normal mod, just that it sends <^>! - shouldn't that work? Maybe an alias mapping AltGr = <^>!
2FIX: The NBSP mapping (AltGr+Spc), in Messenger at least, sends Home or something that jumps to the start of the line?! The first time only, and then normal Space?
2FIX: Remapping to LAlt doesn't quite work? Should we make it recognizeable as a modifier? Trying 'SC038 = LAlt VK' also disabled Extend?
TEST: ToM Ctrl on a letter key? Shift may be too hard to get in flow, but Ctrl on some rare keys like Q or D/H would be much better than awkward pinky chording.
		- It works well! But then after a while it stops working?

2FIX: The HIG doesn't make space between dual accents anymore? They coalesce on AltGr+8 now.
		- Sort of fixed it by making the new disp0 entry that can display any string on the DK's key in the help image.
2FIX: I messed up Gui Show for the images earlier, redoing it for each control with new img titles each time. Maybe now I could make transparent color work? No...?
2FIX: If a layout have fewer states (e.g., missing state2) the BaseLayout fills in empty mappings in the last state! Hard to help? Mark the states right in the layout.
2FIX: Pressing a DK twice should release basechar1 (s1) but basechar0 (s0) is still released. Not sure why.

;;  ================================================================================================================================================
;;  eD TONEXT:

NEXT: Try to emulate AHK Send in such a way that it doesn't send KeyUp even for state-mapped layouts!
		- Just adding " DownR}" to the normal pkl_SendThis() didn't work; the KeyUp events are still sent.
		- Ask around at the AHK forums as to what Send really does, and whether there's an existing workaround for KeyUp. Or at the AHK Discord!
		- One possibility might be to send keys for simple letters, but that's not robust vis-a-vis the OS layout? There's the ## mappings for that, too.

NEXT: Arabic phonetic layout! Also, check the Hebrew one.
		- Hebmak: https://forum.colemak.com/topic/1458-locale-colemak-variants-for-several-countries-the-edreymar-way/#p19971
		- Arabic: ./Layouts/Colemak/Cmk-eD-Ara
		- Write in each ReadMe about the benefits of special DKs - even if you don't care about niqqud/dagesh/etc.

NEXT: Files override?!
		- Maybe just one file to override all files in Files? Maybe place it in root? Have supersections to separate the files.
		- Each supersection could have an entry specifying which file it overrides: _eD_Compose, _eD_DeadKeys, _eD_Extend, _eD_PwrStrings.
		- Not _eD_Remap in this file, as it is already taken care of in the LayStack files?

NEXT: Instead of *etLayInfo("ExtendKey"), an array of mod keys?
		- In the case of more than one, say, SwiSh or Ext keys, could number them? Have each mod entry be an array.
		- { "Extend" : [ "SC###", "SC###" ], "SwiSh" : [ "SC###" ] }, for instance
		- Next up, maybe specify which layer(s) goes which which key so you can have different Extend keys? A dedicated Ext2 key if you want.

NEXT: Update to newer AHK! v1.1.28.00 worked mostly but not for AltGr which sends Alt and gets Ctrl stuck. v1.1.27.07 works fully.
		- AHK version history: "Optimised detection of AltGr on Unicode builds. This fixes a delay which occurred at startup (v1.1.27) or the first Send call (earlier)."
		- After update past v1.1.28, we can use StrSplit() with MaxParts to allow layout variant names with hyphens in them!
		- Should then be able to go to v1.1.30.03 right away, but check for v1.1.31? That version has added an actual switch command, though!!!

NEXT: Add the Canaria variant of Canary, which is said to be much better for Spanish/Spanglish.
		- Compared to standard Canary, it has a J-X swap, AltGr mappings like EsAlt, and grave/tilde DKs (already in eD).
		- https://github.com/christoofar/canaria

NEXT: Rework Extend mappings so they use normal state mapping syntax. The current state of affairs just confuses people.
		- This means that many current mappings that are just `<key>` must be changed to `β{<key>}`.
		- The magic happens in extendKeyPress() in pkl_keypress.ahk, under `if not pkl_ParseSend()`.
		- Here, `Send {Blind}%pref%{%xVal%}` is used after checking for ExtMods.

NEXT: Toggle-type modifiers.
		- Navigating web pages, I'd like to press a lot of PgUp/Dn and arrows without holding the Ext modifier which gets tiresome.
		- I'd also like to stay in the NumPad layer for protracted number entry sessions, without holding down the Ext key.
		- It might be good to have the lock within the layer. Say, Ext+` locks whatever Ext layer is active.
			- For the current Ext+` mapping, it's enough to have that on Ext-tap+`.
			- Or is double-tapping the modifier better, as it doesn't use up a mapping? That, however, would be a problem for ToM keys.
		- But how to make it easy to get out of the layer again? Actually, just tapping the normal modifier should do the trick!
		- Leave lockable layers out of Janitor cleanup? Or just use a long enough Janitor idle timer that it won't be an issue? Configurable?
			- The _pklJanitorCleanup() fn is currently not active, seemingly without any negative impact. So that's okay?

NEXT: Actual settings shown in the Layout Picker and Special Keys tabs.
		- https://github.com/DreymaR/BigBagKbdTrixPKL/issues/80
		- When done, release v1.4.2: Layout/Settings enhancements? Or add some other GUI stuff first (see below)?
NEXT: Add a Help button with a more generic help screen for the first Settings UI panel?
NEXT: Move the text for the Settings UI help text to the language files?!
		- Make a separate .ini file section for it. Then read in the whole section and process it?
NEXT: Flesh out menu entries in the Settings UI? For instance, ANS ⇒ ANS(I), AWide ⇒ AWide (Angle+Wide) etc. Use a dictionary of string replacements?
NEXT: In preparation of AHK v2, rework the Gosub-based routines in the pkl_gui_settings.ahk file.

TODO: Auto-hide help images!? Set a timer for idle time with the Janitor. Inspired by the on-screen keyboard app OverKeys.

TODO: Update the X11 Compose table to a newer version.
		- Must re-import it, and then make a 3-way diff to include all my custom changes to the old one.

TODO: As a `##` state map entry maps to a VK send, maybe add a `#<VK>` syntax to send another VK by its one-letter code or `0x##` code? `#A` for the `A` key, etc.
		- That would encroach upon the AHK syntax, which we've mostly avoided thus far. But `Win+a` could still be easily enough sent by `α#a`, etc.

NEXT: Further getWinLayDKs() development
		- What to do w/ the detect/get/setCurrentWinLayDeadKeys() fns?
		- Get rid of the systemDeadKeys setting, and update setCurrentWinLayDeadKeys() accordingly... unless it's still needed for pkl_Send()?!?
		- Get rid of [DefaultLocaleTxt] and [DeadKeysFromLocID] in EPKL_Tables.ini and all language files?
		- getCurrentWinLayDeadKeys() is checked in pkl_Send(). It's chr based though. Make another dic based on chars, in getWinLayDKs()? But ToAscii doesn't give them?
		- What about pkl_CheckForDKs() in pkl_send.ahk?

NEXT: Bg update according to Kharlamov: Lose duplicate ъ (one on y and one on =+)
		- I think the bulgarian =+ position should house Ѝ ѝ
		- It's a precomposed letter used for homophone distinctions and is present on newer bulgarian layouts
		- Also, there seems to be no ё in bulmak (for russian), even though there's still the russian ы э
		- (The ё could be on `AltGr+/`, since that only houses a duplicate slash and the non-cyrillic ¿)? No, breaks the Latin layer?
NEXT: Belarus/Ukrainia variants? Kharlamov in Mods-n-Layers (messID 961236439591432222 ff):
		- Belarusian can use russian with И и changed to І і, Щ щ changed to Ў ў, and Ъ ъ changed to Ґ ґ
		    (not used in the official orthography, but used in the still-popular 1918 orthography)
		- Russian letters should also be accessible seeing how belarus is officially bilingual
		- The ’ [Cmk-eD AltGr+F] apostrophe too, it's a letter in belarusian
		- The national layout uses `'` so the current mapping may suffice
		- Maybe put ’ on the iso key instead of double acute?
		- For better phonetic mapping, Ў ў should be mapped to W w due to making the same sound

NEXT: A debug hotkey to generate a set of help images on the fly using default settings? Just call the make image fn() then sleep 600 then hit Enter, basically.

NEXT: Since Compose tables can be case sensitive now, do the same for DKs? Then scrap the silly `<K>+`-type DK entry syntax - keep <#> syntax?
		- Read in all DK tables in use at startup instead of each entry as needed then? Faster use, slower startup, more memory usage. Acceptable?

NEXT: Move all override (and settings?) files to the Data folder? More compatible w/ the PortableApps format (backup++), but less clear?

NEXT: User working dir: All user files can be under a specified user root dir.
		- https://github.com/DreymaR/BigBagKbdTrixPKL/issues/34
		- Make a userDir setting in Settings with `Data` as its default value, and read it in during early pkl_init.
			- The Settings_Override specifying where to find user root would have to be at program root.
				- Or? Could use Data as default? Then read it twice, to see if Data holds a Settings_Override. The Data dir could also hold templates?
		- Any file read should look first under user root, and then if not found there program root.
			- Keep tabs in a global array after first access attempt, to avoid looking for files so much?
		- Keep Layout_Override.ini files in their correct layout folder paths but under user root then.
		- Make and look for overrides in the working dir, and defaults in the script dir only?
		- Can also have overriding help images this way. Make the HIG able to generate these under user root?
			- This would make sense, letting a user set individual settings in their working dir and getting images there too.
		- By default, the working dir root can be Data which is the right place for it in the PortableApps standard (for backup).
		- Special syntax? `User\` or `~\` could point to working dir, and `.\` continue to point to script dir? Make all file-reading operations aware of this!?
		- Might use a switch of working dir for some operations? Or, should things like the HIG just assume working dir and anyone wishing to use it must adjust?
			- For users having their EPKL install in a non-writable area, the HIG would need to use the working dir as its work area anyway.

TODO: A layout array to switch layouts, instead of reloading EPKL each time?!
		- The longish delay between full EPKL restarts makes multi-layout usage a challenge
		- Would have to use multi-dimensional dead key etc arrays too, then

TODO: Modifier lock. For which modifiers?
		- Extend lock? Maybe too confusing. But for, say, protracted numeric entry with Ext2 it could be useful?
			- E.g., LShift+Mod2+Ext locks Ext2? Or, have a mapping within the Extend layer that locks it? And one that unlocks all Extend states.
		- Make a more generic lockable modifier type, for SwiSh/FliCK etc (but hardly for Ctrl-lock?).
			- For instance, RCtrl+SwiSh could lock/unlock SwiSh: Hold RCtrl, then tap SwiSh to lock/unlock.
			- There could be a setting for mod-locking pairs? Like this: `SwiSh/RCtrl,LShift/RShift`.
		- Could be a "lock the next modifier" mapping. Especially if we could make that dual-function on, say, the LCtrl key with a ToM timer?
			- If, say, LShift is a mod-lock ToM, then it shouldn't be able to lock itself. Could it be both a sticky Shift and a mod-lock?

TODO: Layout preview image for the layout chooser. Could just change the normal help image temporarily?
		- Show the help image of the chosen layout, with the right background. Allow any state image by pressing modifiers.

TODO: Try to fix AltGr so we can move beyond AHK v1.1.27 which is old now. Ask at the AHK forums!

TODO: Like SteveP's Seniply has it, make Ext-mods Sticky by default? Allow a Parse-Entry syntax for it, or a setting?
		- It's already started in the {Shift OSM} syntax, but not sure that'll work fully with the Ext-mods? They need to be dual-function.

TODO: The newLID pklJanitor routine doesn't quite work, since the locale gets preloaded on EPKL startup or smth. Need to restart EPKL then? Or just re-read parts?

TODO: A layout2/3/4 setting in layout files that can define Swish/Flick layers. Allows for instance a Greek layout added as Swish/Flick layers.

TODO: More GUI settings?
		- A Hotkeys settings panel?
		- Menu language choice (on the Settings tab), with a dropdown choice of the actual language files present?

TODO: If we get Timerless EPKL working, maybe ToM can have better timing at last? Enabling exciting projects like HomeRowMods and a Shift+CoDeKey ToM.
TODO: Once ToUnicode() and DetectDK() are working, it should be possible to generate help images from VK/SC layouts too?!
TODO: SwiSh/FliCK should be the ideal way of implementing mirrored typing?
		- Need to solve them being effectivly one-shot now, then. They should be able to be held down reliably.
		- For fun, could make a mirror layout for playing the crazy game Textorcist: Typing with one hand, mirroring plus arrowing with the other!
TODO: OS DK detection sucks.
		- Go through all SC### and send their four states? (Only if the OS layout has AltGr; can we detect that by DLL?)
		- Or detect through the ToUnicodeEx() DLL? Seems much better.
		- Also store the DK characters in a better format? Just a string like ´¨`^~ is unclear and tricky.
TODO: Ext layers by app/window? Like auto-Suspend. Could be handy for ppl w/ apps using odd shortcuts.
TODO: Look into this Github README template? https://github.com/Louis3797/awesome-readme-template
TODO: Make key presses involving the Win key send VK codes. This'll preserve Win+‹key› shortcuts without using ## mappings.
TODO: Rework the GUI submit to allow multi-submit/reset on tabs that have more than one submit. 
		- Maybe make the submit routine callable with arrays so it loops before asking for a restart?
TODO: IPA Compose sequences, based on my old IPA DK ideas. Vowels with numbers according to position?
TODO: Make a "base compose output" that a Compose key releases whenever no sequence is recognized? Like the Basechar of a DK. Useful for locale layouts?
TODO: Personal override files for extend, compose, powerstrings etc? One override file with sections? Some overrides (remaps, DKs) in layouts.
TODO: Is the main README still too long? Put the layout tutorial in a Layouts README? Also make a tutorial for simply using the CkAWS remap or something.
TODO: Add QWERTZ and AZERTY layouts? There are now remaps for them, and the rest should be doable with OEM VK detection.
TODO: Provide a swap-LAlt-n-Caps RegEdit script, and a reversal one. Maybe add some more codes in the comments, see my old RegEdit scripts.
TODO: Harmonize Ext and folder mod names? And/or make a shorthand for the @E=@C@H@O battery in addition to @K in layout files? And also the short variant like CAW(S)?
		- Could expand, e.g., CurlAWide to CurlAngleWide for the layout name only? Or use long names like CurlAWideSym consistently?
		- Make long names more consistent? Like 4 letters per mod, CurlAnglWideSyms ? Nah, too anal. Better to keep with CurlAWideSym, and that's long enough really.
		- Use CAngle or CA--, etc? CAngle is more intuitive, but CA more consistent with CAW(S). 
TODO: Make a matrix image template, and use it for the Curl variants w/o Angle. 
		- Maybe that should be a separate KbdType, but we also need ANS/ISO info for the VK conversions. ASM/ISM KbdTypes?

;;  ================================================================================================================================================
;;  eD TODO:

TODO: Could the [layout] section be composed from includes of other sections? Such as [Numbers], [Symbols], [Letters], [Others]?
		- That'd be more similar to the modularity of my XKB-data files.
		- This would facilitate hybrid layout types such as VK-numbers (to allow Win+Number shortcuts), VK-letters/eD-symbols...

TODO: Improve GUI responsiveness further.
		- GUI startup is slow. Where are the bottlenecks? Try to tic-toc GUI startup?
		- Globals initialization including folder scanning has been added to a function called at init time.
		- Could also make GUI startup more responsive by pre-initializing the GUI at init. As it is now, it's made from scratch each time.
		- Candidate: Layout selection. Premake the list of eligible layouts (and LayType? KbdType? Variants? Mods?) beforehand.
TODO: Use the off-Space thumb key as a Shift/CoDeKey ToM!
		- Preserves SteveP's thumb Shift for compact boards, while allowing the fancy CoDeKey shift combos too.
		- It's a sweet idea! But again, it requires better ToM timing than EPKL can currently deliver. So it'll have to wait, for now...
TODO: Make the CoDeKey follow the StickyTime timer? So you'll only use it as CoDeKey in flow. No, it'd need its own timer.
TODO: Could I turn around the Compose method, to be leader key after all? But how to input then? Without looking sucks. In a pop-up box?
TODO: Color markings for keys in HIG images! Could have a layer of bold key overlays and mark the keys we want with colors through entries in the HIG settings file.
		- markColors = #c00:_E/_N/_K, #990:_B/_T/_F, #009:_J     ; Tarmak2 colors
		- markColors = <CSV of marking specs>, similar to the remaps. Could have Tarmak1,Tarmak2,Tarmak3,#009:_J ?
		- See https://forum.colemak.com/topic/1858-learn-colemak-in-steps-with-the-tarmak-layouts/p4/#p23659
		- Allow a section in Layout.ini too!
		- Mark differently by state, as in the Tarmak images
TODO: Make state images and DK image dirs ISO/ANSI aware?! Generate both in the HIG each time (plus Ortho?). Make layouts that can handle both. 
		- How to handle special mappings? Could have [layout_###] sections.
TODO: I never use the SendMessage parse prefix. Cannibalize it for a strEsc() send? Or add that as €\ prefix instead?
Mod ensemble: For lr in [ "", "L", "R" ], For mod in [ "Shift", "Ctrl", "Alt", "Win" ] ? May not always need the empties? Also add [ "CapsLock", "Extend", "SGCaps" ] ?
TODO: Redo the @Ʃ_@Ç formalism, adding @K to @E(@C@H@O) by a hyphen instead of an underscore? Would that be a benefit in any way? Or just a lot of work?
TODO: Hotstrings? May have to wait for AHK v1.1.28 to use the Hotstring() fn? Or is there somewhere in this script we could insert definitions?
TODO: Consider a remap for each Ext layer? Would make things messier, but allows separate Ext1 and Ext2 maps, e.g., for the SL-BS switch.
		- Allow mapSC_extend2 etc entries in the LayStack. If not specified, use the _extend one for all.
TODO: Add ABNT keys to the HIG template?
TODO: Record macro? Or just a way to set entries for a certain DK layer in the Settings UI? Say, the Ext-tap layer(s). Could have backup DK layers and a Reset button.
TODO: Make EPKL able to hold more than one layout in memory at once?! This would make dual layouts smoother, and using layouts as layers (Greek, mirroring etc) possible.
		- With SGCaps modifier layers, the need for this may be alleviated?
TODO: Since no hotkeys are set for normal key Up, Ext release and Ext mod release won't be registered? Should this be remedied?
TODO: Rework the modifier Up/Down routine? 
		- A function pklSetMods( set = 0, mods = [ "mod1", "mod2", ... (can be just "all")], side = [ "L", "R" ] ) could be nice? pkl_keypress, pkl_deadkey, in pkl_utility
TODO: Replace today's AltGr handling with an AltGr modifier. You'd have to map, e.g., RAlt = AltGr Modifier, but then all the song-and-dance of today would be gone.
		- Note that we both need to handle the AltGr EPKL modifier and whether the OS layout has an AltGr key producing LCtrl+RAlt on a RAlt press.
		- Also allow ToM/Sticky AltGr. Very very nice since AltGr mappings are usually one-shot.
		- Define a separate AHK hotkey for LCtrl+RAlt (=AltGr in Windows)? That might make things simpler.
TODO: VK mappings don't happen on normal keys. Simple VK code states don't get translated to VK##. Only used when the key is VK mapped.
TODO: Instead of CompactMode, allow the Layouts_Default (or _Override) to define a whole layout if desired. Specify LayType "Here" or suchlike?
		- At any rate, all those mappings common to eD and VK layouts could just be in the Layouts_Default.ini file. That's all from the modifiers onwards.
TODO: Import KLC. Use a layout header template.
		- Could have a section of RegEx conversions with name tags in the template, which gets used and then cut out.
		- Each such entry could have a tagName = ## SplitBy JoinBy <regex>
			- Allow both RegExReplace and RegExMatch entries? The latter should use O) match objects?
			-  The ## denotes how many numbered entries should be run on this string. This could have sublevels, like ##-##-##.
			- SplitBy loops through elements of the string, recursively if subentries also split. Then it's rejoined with JoinBy (necessary, or just regex that?).
			- Can we SplitBy words, like \nDEADKEY\t ?
		- Then in the template there's something like $$tagName$$ where the result is to be inserted.
		- For DK full names, the KEYNAME_DEAD entries could be converted (cut out ACCENT/SIGN, _ for spaces?, cut away parentheses, title case). Update my names accordingly?
		- In addition to MSKLC format, allow Aldo Gunsing's KLFC! https://github.com/39aldo39/klfc And maybe Keyboard Layout Editor's KLE (or do that via KLFC).
TODO: Make pklParseSend() work for DK chaining (one DK releases another)!
		- Today, a special DK entry will set the PVDK (DK queue) to ""; to chain dead keys this should this happen for @ entries?
		- Removing that isn't enough though? And actually, should a dk chaining start anew? So, replicate the state and effect of a normal layout DK press.
		- Chaining DKs opens up for interesting possibilities, like a Mother-of-Dead-Keys key (MoDK)! Could that be on Extend-tap, possibly with a timeout? Or on Backspace?
			- See Jaroslaw's MoDK topic in the Forum: https://forum.colemak.com/topic/2501-my-current-programming-symbols-layout/#p22527
			- Placing all my DKs on MoDK sequences will fill up a layer. So maybe only the most interesting ones? But how to make it mnemonic?
			- Example: Tap-dance {Ext,t,n} -> ñ; {Ext,a,A} -> Á; {Ext,0-9} IPA DKs.
			- For good measure, could have different DKs on different states of the same key! Wow. The ToM formalism should support this actually!
TODO: With the Compose method, look into IME-like behavior?
		- This would allow "proper" Vietnamese, phonetic Kyrillic etc layouts instead of dead keys which work "the wrong way around".
		- Could make special compose keys for accents? E.g., you type a^ and the ^ key is a Compose key producing â.
TODO: Make EPKL work with the .exe outside a .zip file? 
		- You could then download the release .zip, put the .exe outside, change then rezip any settings you want to, then the .exe will use the archive.
		- This may be desirable for people running EPKL from an URL. It's easier to handle two files than several folders.
TODO: Try out <one Shift>+<other Shift> = Caps? How to do that? Some kind of ToM, where the Shift is Shift when held but Caps when (Shift-)tapped?
TODO: The key processing timers generate autorepeat? Is this desirable? It messes with the ToM keys? Change it so the hard down sends only down and not down/up keys?
TODO: Keylogging for gathering typing stats. Which stats? 1-2-3-grams, characters-before-backspace...
TODO: A help fn to make layout images? Make the image large and opaque, then make a screenshot w/ GIMP and crop it. Or can I use the Windows Snipping Tool (Win+Shift+S)?
TODO: AHK2Exe update from AutoHotKey v1.1.26.1 to v1.1.30.03 (released April 5, 2019) or whatever is current now. 	;eD WIP: Problem w/ AltGr?
		- New Text send mode for PowerStrings, if desired. Should handle line breaks without the brkMode setting.
TODO: Make the Japanese layout now, since dead keys support literals/ligatures and DK tables in Layout.ini are possible.
TODO: Mirrored one-hand typing as Remap, Extend or other layer?
		- For Extend, would need a separate Ext modifier for it? E.g., NumPad0 or Down for foot or right-arm switching. But is that too clunky?
		- SGCaps could work, but would require each layout to have SGC mappings to allow mirroring then. And a separate SGC modifier.
		- Layout switching is usually done by restarting EPKL which is too clunky. But if we could have a switch modifier that temporarily activates the next layout...?
		- This would require preloading more than one layout which takes a bit of reworking. Possibly... Allow an alt-set of the remap only, remapping on the fly w/ a mod?
		- Mirroring as a remap can now use minicycles of many two-key loops. For instance, |  QU |  SC /  MN |  SL | for two separate swaps.
TODO: A set of IPA dk, maybe on AltGr+Shift symbol keys? Could also be chained from a MoDK?
TODO: Lose CompactMode from the Settings file. The LayStack should do it.
		- Instead of a setting in Settings, allow all of the layout to reside in EPKL_Layouts_Default (or Override). If detected, use root images if available.
		- If no Layout.ini is found, give a short Debug message on startup explaining that the root level default/override layout, if defined, will be used. Or just do it?

;;  ================================================================================================================================================
;;  eD ONHOLD:

HOLD: More `¢[]¢` syntax?
		- Use Eval() on arg; will that allow for instance "str" to be read as a string without the quotes?
			- No, eval() doesn't exist, and its principle is regarded evil in the AHK/coding communities.
		- Add  RunWait()? No: That waits until the program exits, which we don't want.
		- With the new power of Run(), add more handy Ext-tap mappings? Which would be most handy?
		- https://superuser.com/questions/217504/is-there-a-list-of-windows-special-directories-shortcuts-like-temp
		- https://www.autohotkey.com/docs/v1/misc/CLSID-List.htm
		- Could I use dynamic fn calling for wrapper functions in exec()? The original AHK example uses Goto labels (bad), but dynamic fn() could work?
			- https://www.autohotkey.com/boards/viewtopic.php?t=75956

HOLD: Default positional special DK mappings are messed up for non-Cmk layouts. Should there be a DK remap possibility?
		- This affects primarily the CoDeKey (@co0) and Ext-tap (@ex0/1) layers.
		- This issue is layout-based, making (Base-)Layout files the right place to override the mappings - for now.
		- At some point, a proper Remap option for positional DK layers would be neat.

HOLD: For the System layout having state help images makes no sense. Remedy this?
		- Use LayInfo("shiftStates")? But atm, not having the shift states active ruins OS DKs.
		- Cool idea: Make the Vim Help Sheet for Colemak available as a state image? Have it, e.g., on state1 to show it whenever Shift is pressed.
		- Did this, using my colemak-vim-helpsheet.svg files.
		- Ideally, should have different images depending on ergo mods. At least, ISO/ANS -(A)-- + CA-- + CAWS.
		- Problem: The smallest text on the help image doesn't render well at the standard help image resolution.

HOLD: Make WideSym into a separate remap now? Simpler? Less confusing, or not?
		- The Wide mod is already split by row (except RB and BS) which is quite instructive and consistent.
		- Made WS remaps for alt layouts like Sturdy that shouldn't do pure Wide mods as they have stuff on the Slash key.
		- We already have `SymMnW_@K` (no Cmk-QU remap) for alt layouts (vs `SymQuMnW_ANS`, for Cmk etc.).
		- An 'UnSym' remap is a good way of resetting symkey mappings for some layouts (like Graphite) before proceeding.

HOLD: Even More Modern Alt Keyboard Layouts?
		- We already have Semimak-JQ (2021) and Canary (2022), so consider adding some other "best candidates".
			- https://getreuer.info/posts/keyboards/alt-layouts/index.html#which-alt-keyboard-layout-should-i-learn
			- APT, Nerps, Sturdy and maybe Engram look good in Getreuer's comparison. [Also, Gallium and Graphite now, ++.]
		- APTv3 by Apsu (2021; the currently stable version) has been added.  https://github.com/Apsu/APT
			- Also add Aptmak? It uses a thumb key; EPKL could handle that (maybe w/ help from SharpKeys).
		- Graphite (2022-12) by RDavison was also added after favorable mentions.  https://github.com/rdavison/graphite-layout
		- Sturdy by Oxey (2022)?  https://o-x-e-y.github.io/layouts/sturdy/index.html
			- Oxey has several candidates, but I'd focus on one. It has high rolls but still low redirects.
			- Oxey said that Sturdy has a "decent amount of users" but "people are dropping it" so not worth it now?
			- https://discord.com/channels/409502982246236160/1002128319770271834/1129455798368612484
			- Con: Like Nerps, it has a too common key – CM – in the SL position. However, Oxey had some ideas for Wide.
				- https://www.reddit.com/r/KeyboardLayouts/comments/15zu2rn/comment/jysbcn8/?context=2
			- Possibly also Magic Sturdy by Ikcelaks, which EPKL could do...? It's what Getreuer uses.
			- https://github.com/Ikcelaks/keyboard_layouts/blob/main/magic_sturdy/magic_sturdy.md
		- Nerps by Smudge (2022)? I don't like its PD(.<) in the SL(/?) key position: Messes with the Wide mod, and just feels odd.
			- Also, it seems unstable. According to Oxey, it's been "superseded by Nerbs and Gallium" in some respects. No home page.
		- Engram?  https://engram.dev/  by Arno Klein (2021) is also neat, and some recommended adding it. Higher SFB% than CMK though (corpus dependent)!?
		- Dwarf/Whorf? ("Parented to CMini"?). No home pages?

	- Allow a mapping like Modifier(#), to add # to the modifier level? Use it as single-argument mapping entry. Modifier(8) would be SwiSh.
	- Instead of having to make special literal entries (`→` or similar) for unshifted characters in shifted states, make all character sends use Unicode/Text?
		- Issue: With Sticky Shift, the 2nd state mapping is sent shifted which is wrong if it was mapped to be something unshifted. Normal Shift does not.
		- Sticky Shift just holds down the Shift key which leads to this effect. Should I make sure the state map is sent unblind?
		- Only happens for single-character mappings. Mappings that aren't a key name aren't sent as "keys" by AHK.
		- Conclusion: Not a good idea to send as text categorically, as non-"key" sending breaks Win+‹key› shortcuts.
	- Try out a swap-side layout instead of the mirrored one? More strain on weak fingers, but fewer SFBs I should think.
		- Is the brain equally good at side-swapping and mirroring?
	- Make @K a compound (ANS/ISO-Trad/Orth/Splt/etc)? ANS/ISO is needed for VK codes, and the form factor for images and layout subvariants. kbdType vs kbdForm?
		- Could keep everything in kbdType and adjust the reading of it to use the first and second substring.
		- However, it may not be necessary at all. Using a kbdType like ANS-Orth seems to work just fine for now. The VK-related kbdType is in Layout.ini anyway.
	- Hardcode Tab instead of using &Tab after all? It's consistent to have both the whitespace characters Spc & Tab hardcoded this way.
	- A dynamic key press indicator for help images, showing not just modifier layer but every press. Will it be fast enough? Needs a position table for each KbdForm.
	- Make a Setting for which fn to run as Debug, so I don't have to recompile to switch debug fn()? Maybe overmuch, as the debug fn often needs recompiling anyway?
	- Allow Remaps to use @K so that the layouts don't have to?!? Too confusing?
	- Remove all the CtrlAltIsAltGr stuff? If laptops don't have RAlt (>!), they can just map a key to AltGr Mod instead? Won't allow using <^<! as AltGr (<^>!) though...
	- Shift sensitive multi-Extend? When mapping for the NumPad layer, it'd be nice to have $/¢, €/£ etc. This allows many more potential mappings! 4×4-level Extend?!
		- In most cases though, that'd be useful mostly for releasing more different glyphs. This is better done with dead keys, as these avoid heavy chording.
	- Allow escaped semicolons (`;) in iniRead?
	- Remove the Layouts submenu? Make it optional by .ini?
	- Greek polytonic accents? U1F00-1FFE for circumflex(perispomeni), grave(varia), macron, breve. Not in all fonts! Don't use oxia here, as it's equivalent to tonos?
	- Some kaomoji have non-rendering glyphs, particularly eyes. Kawaii (Messenger), Joy face, Donger (Discord on phone). Just document and leave it at that.
	- Go back on the Paste Extend key vs Ext1/2? It's ugly and a bit illogical since the layers are otherwise positional. But I get confused using Ext+D for Ctrl+V.
	- Allow assigning several keys as Extend Modifier?
	- An EPKL sample Layout.ini next to the original PKL one, to illustrate the diffs? Or, let the contents of the main README be enough?
	- Auto language detection doesn't follow keyboard setup but system language. If you use a Non-English keyboard but Windows uses English, the auto language is English.
	- Someone used CapsLock on double-tap Shift. Neat idea! Except for the double-tap SFB, of course. If you have both Shift keys, a chord could do it?!

;;  ########################   <--«  TODO   ########################
*/

;;  ########################    main code   ########################

;#Requires AutoHotkey 1.1.27     							; This will be important when transitioning to AHK v2 in the future. At that point, use `2.0+` here.
;#Warn                   									; This AHK directive is handy for spotting global/local variable conflicts etc. May give a lot of warnings though.
#NoEnv
#Persistent
#NoTrayIcon
#InstallKeybdHook
#SingleInstance         Force   							; eD WIP: Is something wonky with this now? I get lots of apparent EPKL instances in the System Tray...?
#MaxThreadsPerHotkey    3
#MaxThreadsBuffer       Off 								; We'll turn it on later in pkl_init, so it's off for program hotkeys and on for key press ones.
#MaxHotkeysPerInterval  300
#MaxThreads             32
#MaxMem                 128 								; Default 64 Mb. We need more than that for HIG image generation in its search-n-replace loop.
#KeyHistory             24  								; NOTE: If you're concerned about the security risk of the AutoHotKey KeyHistory log, set this to 0 and recompile.

SendMode Event
SetKeyDelay,    -1  										; The Send key delay wasn't set in PKL, defaulted to 10. AHK direct key remapping uses -1. What's most robust?
SetBatchLines,  -1  										; This script never sleeps (default is every 10 ms)
Process, Priority, , R  									; Real-time process priority (default is N for Normal; H for High in old PKL; R for Realtime is max)
SetWorkingDir,  %A_ScriptDir%   							; Should "ensure consistency" 	; eD WIP: Make a separate user working dir, so users can have their settings elsewhere?
StringCaseSense, On 										; All string comparisons are case sensitive (AHK default is Off) 	; eD WIP: But InStr() is still caseless by def.?

setPklInfo( "pklName", "EPiKaL Portable Keyboard Layout" )  				; EPKL Name
setPklInfo( "pklVers", "1.4.2α" )       									; EPKL Version
setPklInfo( "pklHome", "https://github.com/DreymaR/BigBagKbdTrixPKL" )  	; URL - used to be http://pkl.sourceforge.net/
setPklInfo( "pklHdrA", ";`r`n;;  " ) 										; A header used when generating EPKL files  	; eD WIP: Import Module
setPklInfo( "pklHdrB", "`r`n"
		. ";;  for EPiKaL Portable Keyboard Layout (EPKL) by Øystein ""DreymaR"" Bech-Aase (2015-), based on PKL by Máté Farkas (2008-2010).`r`n"
		. ";`r`n" )

setPklInfo( "initStart", A_TickCount )  					; eD DEBUG: Time EPKL startup
;;  Global variables are now largely replaced by the get/set info framework, and initialized in the init fns
;global UIsel   											; Variable for UI selection (use Control names to see which one) 	; NOTE: Can't use an object variable for UI (yet)
Gosub setUIGlobals  										; Set the globals needed for the settings UI (is this necessary?)
initPklIni( A_Args[1] ) 									; Read settings from pkl.ini (now PklSet and PklLay). Get layout from command line parameter (legacy %1%), if any.
initLayIni()    											; Read settings from Layout.ini and layout part files
activatePKL()
;pklDebug( "Time since init start: " . A_TickCount - getPklInfo( "initStart" ) . " ms", 1 ) 	; eD DEBUG

Return  													; <--«  main code

;;  ########################    labels      ########################

;;  eD WIP: Map AltGr to RAlt to prevent trouble?!
;;  The order of AltGr is always LCtrl then RAlt. Custom combos always have the * (wildcard) mod so they obey any mod state.
;;  In order to make a combo hotkey for LCtrl&RAlt, we also need to handle the first key on its own (https://www.autohotkey.com/docs/Hotkeys.htm#combo)
;;  "For standard modifier keys, normal hotkeys typically work as well or better than "custom" combinations. For example, <+s:: is recommended over LShift & s::."
;;  Possible issue: These hotkeys are generated after the others, since initPklIni() is already run. Should this part be handled in the init part? What about any LCtrl hotkey in the layout?
;#if GetKeyState( "LCtrl", "P" )
;RAlt::
;#if
;LControl & RAlt:: 	; This works but mapping to RAlt produces "Invalid hotkey", why!? Also, it repeats.
;<^>!:: 				; eD WIP: This isn't working?! Maybe an #if GetKeyState( "RAlt", "P" ) will do the trick?
;	pklDebug( "Gotcha, AltGr!", 0.5 )
;Return
;LControl & RAlt::Send {RAlt Down} 	; This alone gets AltGr stuck
;LControl Up & RAlt Up::Send {RAlt Up} 	; This doesn't work!?

keypressDown: 			; *SC###    hotkeys
	Critical
	keyPressed(         SubStr( A_ThisHotkey, 2     ) ) 	; SubStr removes leading '*'
Return

keypressUp:  			; *SC### UP
	Critical
	Send % "{Blind}{" . getKeyInfo( SubStr( A_ThisHotkey, 2, -3 ) . "ent1" ) . "  UP}"  	; Send the remapped key up
Return

modifierDown: 			; *SC###    (call fn as HKey to translate to modifier name)
	Critical
	setModifierState( getKeyInfo( SubStr( A_ThisHotkey, 2     ) . "ent1" ), 1 )
Return

modifierUp:
	Critical
	setModifierState( getKeyInfo( SubStr( A_ThisHotkey, 2, -3 ) . "ent1" ), 0 )
Return

tapOrModDown: 			; *SC###
	Critical
	setTapOrModState(   SubStr( A_ThisHotkey, 2     ), 1 )
Return

tapOrModUp:
	Critical
	setTapOrModState(   SubStr( A_ThisHotkey, 2, -3 ), 0 )
Return

showAbout: 													; Menu "About..."
	pkl_about()
Return

changeSettings: 											; Menu "Layout/Settings..."
	pklSetUI()
Return

keyHistory: 												; Menu "AHK Key History..."
	KeyHistory
Return

;detectCurrentWinLayDeadKeys: 								; Menu "Detect dead keys..."
;	setCurrentWinLayDeadKeys( detectCurrentWinLayDeadKeys() )
;Return

showHelpImage:
	pkl_showHelpImage()
Return

showHelpImageOnce: 											; Used as a one-time refresh when necessary
	pkl_showHelpImage()
Return

toggleHelpImage:    										; Menu "Display help image"
	pkl_showHelpImage( 2 )
Return

zoomHelpImage: 												; Menu "Zoom help image"
	pkl_showHelpImage( 5 )
Return

moveHelpImage: 												; Hotkey "Move help image"
	pkl_showHelpImage( 6 )
Return

opaqHelpImage: 												; Hotkey "Opaque/Transparent image"
	pkl_showHelpImage( 7 )
Return

rerunNextLayout: 											; Menu "Change layout"
	changeLayout( getLayInfo( "NextLayout" ) )
Return

rerunSameLayout:    										; Menu "Refresh EPKL"
	activeLay   := getLayInfo( "ActiveLay" ) 				; Layout code (path) of the active layout
	numLayouts  := getLayInfo( "NumOfLayouts" ) 			; The number of listed layouts
	Loop % numLayouts { 									; Use the layout # instead of its code, to reflect any PKL Settings list changes
		theLayout   := getLayInfo( "layout" . A_Index . "code", theCode )
		actLayNum   := ( theLayout == activeLay ) ? A_Index : actLayNum
	}
	changeLayout( "UseLayPos_" . actLayNum ) 				; Rerun the same layout, telling pkl_init to use position.
Return

changeLayoutMenu: 											; Menu "Layouts"
	changeLayout( getLayInfo( "layout" . A_ThisMenuItemPos . "code" ) )
Return

suspendOn:
	Suspend, On
	Goto afterSuspend
Return

suspendOff:
	Suspend, Off
	Goto afterSuspend
Return

toggleSuspend:  											; Menu "Suspend"
	Suspend
	Goto afterSuspend
Return

afterSuspend:
	If ( A_IsSuspended ) {
		pkl_showHelpImage(  3 )
		Menu, Tray, Icon, % getLayInfo( "Ico_OffFile" ), % getLayInfo( "Ico_OffNum_" )
	} else {
		pkl_showHelpImage( -3 )
		Menu, Tray, Icon, % getLayInfo( "Ico_On_File" ), % getLayInfo( "Ico_On_Num_" )
	}
Return

exitPKL: 													; Menu "Exit"
	ExitApp
Return

doNothing:
Return

openTarget:
	runTarget( getPklInfo( "openMenuTarget" ) ) 		; Open the chosen target (default A_ScriptDir, in File Explorer)
Return

getWinInfo:
	getWinInfo() 										; Show the active window's title/process(exe)/class
Return

epklDebugUtil:  										; eD DEBUG/UTILITY/WIP: This entry is activated by the Debug hotkey
	nr  := pklIniRead( "whichUtility", 1 )
	pklToolTip( "Running Debug/Utility routine " . nr . "`n(specified in Settings)", 1.5 )
	debug%nr%() 										; Run the specified debug# routine
Return

  debug1() {
	KeyHistory  										; Show AHK Key History      as by the View -> Key history menu  (shown)
} debug2() {
	ListHotkeys 										; Show AHK hotkeys          as by the View -> Hotkeys     menu  (hidden)
} debug3() {
	ListVars 											; Show AHK global variables as by the View -> Variables   menu  (hidden)
} debug4() {
	ListLines   										; Show AHK script (flow relevant) line execution history        (hidden)
} debug5() {
	getWinInfo() 										; Show the active window's title/process(exe)/class             (EPKL)
} debug6() {
	menuIconList()  									; List menu icons in a specified file
} debug7() {
	pklDebugCustomRoutine() 							; eD DEBUG – usually: Show OS & EPKL VK codes for the OEM keys
} debug8() {
	detectCurrentWinLayDeadKeys()   					; The old PKL DeadKey detection routine                         (hidden)
} debug9() {
	getWinLayDKs()  									; eD WIP: Improved WinLayDK detection
	pklDebug( "getWinLayDKs:`n" . getPklInfo("WinLayDKs")[0x10], 2 )  ; eD DEBUG
;	importLayouts() 									; eD TODO: Import a MSKLC layout file to EPKL format
;	importComposer() 									; eD DONE: Import an X11 Compose file to EPKL format
}   ; <-- debug#

;;  ########################    includes    ########################

#Include pkl_init.ahk
#Include pkl_gui_image.ahk      	; pkl_gui was too long; it's been split into help image and menu/about parts
#Include pkl_gui_menu.ahk
#Include pkl_gui_settings.ahk
#Include pkl_keypress.ahk
#Include pkl_send.ahk
#Include pkl_deadkey.ahk
#Include pkl_utility.ahk        	; Various functions such as pkl_activity.ahk were merged into this file
#Include pkl_get_set.ahk
#Include pkl_ini_read.ahk
#Include pkl_import.ahk         	; Import module, converting MSKLC layouts to EPKL format, and other import/conversion
#Include pkl_make_img.ahk       	; Help image generator, calling Inkscape with an SVG template

;;  ########################    functions   ########################

;;  Functions and further subroutines are defined in the included files.