tynet-lichat/shirakumo

17:47:21 shinmera scymtym: hitting an error when defining a text shape with an extent rather than a margin:

17:47:29 shinmera There is no applicable method for the generic function #<STANDARD-GENERIC-FUNCTION (COMMON-LISP:SETF ORG.SHIRAKUMO.ALLOY::MAX-WIDTH) (4)> when called with arguments (16.0 NIL).

17:47:43 shinmera 2: ((:METHOD ORG.SHIRAKUMO.ALLOY:SUGGEST-SIZE (T ORG.SHIRAKUMO.ALLOY.RENDERERS.OPENGL.MSDF:TEXT)) (ORG.SHIRAKUMO.ALLOY:EXTENT 2.0un 1.0un 16.0un 16.0un) #<ORG.SHIRAKUMO.ALLOY.RENDERERS.OPENGL.MSDF:TEXT b.. 3: ((:METHOD SHARED-INITIALIZE :AFTER (ORG.SHIRAKUMO.ALLOY.RENDERERS.OPENGL.MSDF:TEXT T)) #<ORG.SHIRAKUMO.ALLOY.RENDERERS.OPENGL.MSDF:TEXT bounds (ORG.SHIRAKUMO.ALLOY:EXTENT 2.0un 1.0un 16.0un 16.0un) re.. 4: ((SB-PCL::EMF SHARED-INITIALIZE) #<unused argument> #<unused argument> #<ORG.SHIRAKUMO.ALLOY.RENDERERS.OPENGL.MSDF:TEXT bounds (ORG.SHIRAKUMO.ALLOY:EXTENT 2.0un 1.0un 16.0un 16.0un) real size NIL {121.. 5: ((:METHOD MAKE-INSTANCE (CLASS)) #<STANDARD-CLASS ORG.SHIRAKUMO.ALLOY.RENDERERS.OPENGL.MSDF:TEXT> :TEXT "" :BOUNDS (ORG.SHIRAKUMO.ALLOY:EXTENT 2.0un 1.0un 16.0un 16.0un) :FONT #<ORG.SHIRAKUMO.ALLOY.RE.. 6: ((LABELS SB-PCL::CALL-NEXT-METHOD-N :IN "/home/linus/Projects/cl/alloy/renderers/simple/defaults.lisp") #<ORG.SHIRAKUMO.FRAF.TRIAL.IDE::UI UI> (ORG.SHIRAKUMO.ALLOY:EXTENT 2.0un 1.0un 16.0un 16.0un) "".. 7: ((:METHOD ORG.SHIRAKUMO.ALLOY.RENDERERS.SIMPLE.PRESENTATIONS:REALIZE-RENDERABLE (ORG.SHIRAKUMO.ALLOY.RENDERERS.SIMPLE.PRESENTATIONS:DEFAULT-LOOK-AND-FEEL ORG.SHIRAKUMO.ALLOY:MENU-ITEM)) #<ORG.SHIRAKUM.. 8: ((SB-PCL::EMF ORG.SHIRAKUMO.ALLOY.RENDERERS.SIMPLE.PRESENTATIONS:REALIZE-RENDERABLE) #<unused argument> #<unused argument> #<ORG.SHIRAKUMO.FRAF.TRIAL.IDE::UI UI> #<ORG.SHIRAKUMO.ALLOY:MENU-ITEM Open (.. 9: ((:METHOD ORG.SHIRAKUMO.ALLOY.RENDERERS.SIMPLE.PRESENTATIONS:REALIZE-RENDERABLE :AROUND (ORG.SHIRAKUMO.ALLOY.RENDERERS.SIMPLE.PRESENTATIONS:RENDERER ORG.SHIRAKUMO.ALLOY.RENDERERS.SIMPLE.PRESENTATIONS:.. 10: ((:METHOD ORG.SHIRAKUMO.ALLOY:REGISTER :AROUND (ORG.SHIRAKUMO.ALLOY:LAYOUT-ELEMENT ORG.SHIRAKUMO.ALLOY:RENDERER)) #<ORG.SHIRAKUMO.ALLOY:MENU-ITEM Open (ORG.SHIRAKUMO.ALLOY:EXTENT 0.0px 0.0px 0.0px 0.0..

17:53:31 shinmera also once again hitting the thing where I'd like to use layouts to position shapes

17:53:35 shinmera echusa

17:53:41 shinmera enough for today

18:02:49 shinmera but of course my brain can't turn off

18:03:06 shinmera wondering if we can somehow re-use layouts for shape positioning and sizing strategy

18:43:09 Colleen Shinmera: should be fixed in my version. i can probably push soon. passing the :wrap initarg fixed a lot of stuff

18:43:09 scymtym Shinmera: should be fixed in my version. i can probably push soon. passing the :wrap initarg fixed a lot of stuff

18:43:54 Colleen but besides that, i'm not sure whether my changes (to use the new text info for everything) break the toolbar and editor tabs

18:43:54 scymtym but besides that, i'm not sure whether my changes (to use the new text info for everything) break the toolbar and editor tabs

18:44:26 Colleen namely, the toolbar buttons have no icons or anything. is that expected?

18:44:26 scymtym namely, the toolbar buttons have no icons or anything. is that expected?

18:46:04 shinmera they should if you're on the latest of trial, trial-assets, and trial-ide

18:46:19 shinmera (I recompiled the dist with the necessary changes yesterday)

18:47:58 Colleen must be trial-assets. i don't think i updated that

18:47:58 scymtym must be trial-assets. i don't think i updated that

18:48:34 shinmera looks like this for me https://filebox.tymoon.eu/4EFDGXPOQ

18:48:46 shinmera I'll fix all the styling issues in due time

18:53:11 Colleen it's pretty close after updating trial-assets: https://filebox.tymoon.eu/4EHBHNVT6

18:53:11 scymtym it's pretty close after updating trial-assets: https://filebox.tymoon.eu/4EHBHNVT6

18:53:47 shinmera not sure where that sizing difference is coming from for the buttons

18:54:00 Colleen me neither

18:54:00 scymtym me neither

18:54:18 shinmera they are margin'd labels with v/halign center

18:55:13 shinmera for the menu items not being weirdly wide, that's a recent change I just pushed to the alloy branch

18:57:27 Colleen haha, ok. i ignore the font size, i think

18:57:27 scymtym haha, ok. i ignore the font size, i think

18:57:54 shinmera that'd explain why changing the base size also doesn't change the editor

18:58:05 shinmera *base scale

18:58:19 Colleen yeah. should be a surface-level problem. i just didn't know where to get the value

18:58:19 scymtym yeah. should be a surface-level problem. i just didn't know where to get the value

18:58:50 Colleen i will try to push my changes tomorrow

18:58:50 scymtym i will try to push my changes tomorrow

18:58:57 shinmera thanks!

19:52:56 shinmera I also recently had an idea on how to make finalizers not suck

19:53:30 shinmera namely that you make the finalizer push the foreign pointer and class onto a queue, and then process that queue whenever you feel like it

19:53:44 shinmera the class lets you use a prototype to do the freeing dispatch off of

19:53:55 shinmera and being on a queue lets you process it from the correct thread, at a time that is suitable

19:54:32 shinmera sure, you won't be able to trigger the finalization before gc pressure arrives or whatever, but at least it annulls most of the issues

19:54:52 shinmera and for games at least you can trigger the gc during load

19:56:03 shinmera might take that and other lessons I've learned doing a billion wrappers and turn it into another library

21:10:46 Colleen Speaking of wrappers, it seems pretty odd to me that no one has written a C parser in CL and used it to generate cffi bindings

21:10:46 Visen Speaking of wrappers, it seems pretty odd to me that no one has written a C parser in CL and used it to generate cffi bindings

21:11:09 shinmera there's been multiple projects that have done that :)

21:11:39 Colleen Which? I know of none, all the existing ones I've found have been using an llvm wrapper or something to parse the C code

21:11:39 Visen Which? I know of none, all the existing ones I've found have been using an llvm wrapper or something to parse the C code

21:12:04 shinmera I remember hearing of two, but I don't remember either ever being like, finished, or at least finished to a point where they found use elsewhere

21:12:48 Colleen hmm, cl-bindgen and claw are the two binding generators I know of, and they both rely on external C/C++ libraries/programs to function

21:12:48 Visen hmm, cl-bindgen and claw are the two binding generators I know of, and they both rely on external C/C++ libraries/programs to function

21:13:09 shinmera iirc scymtym did some work like that for one, and the other one that iirc even implemented all the actual call routines and so on to be able to call out with struct values and so on

21:13:21 shinmera can't recall who it was, but it was a #commonlisp regular

21:14:42 shinmera iirc it was called "noffi" but I can't find that on the useless search engines, so

21:15:11 Colleen Oh I found scymtym's, seems to be abandoned https://github.com/scymtym/language.c

21:15:11 Visen Oh I found scymtym's, seems to be abandoned https://github.com/scymtym/language.c

21:15:11 shinmera ah, there we go

21:15:24 shinmera it was gilberth, and indeed https://lispcafe.org/fossil/noffi/home

21:16:10 shinmera anyway, because of ~reasons~ it is not easy to "just" do what you propose

21:16:39 Colleen Oh nice, that one seems to still be developed

21:16:39 Visen Oh nice, that one seems to still be developed

21:17:20 shinmera scymtym probably has about as many projects as I do :)

21:17:41 Colleen Yeah I know C can be tricky to parse, and the preprocessor makes things more difficult, etc..etc.., but I just really miss the seamless ffi experience that zig has when using CL

21:17:41 Visen Yeah I know C can be tricky to parse, and the preprocessor makes things more difficult, etc..etc.., but I just really miss the seamless ffi experience that zig has when using CL

21:18:21 shinmera well zig just uses llvm, no? :v

21:20:16 Colleen Yes but llvm is built into the compiler, so the experience is seamless. I don't need to install anything special to use the bindings generator. Also I think zig uses this https://github.com/Vexu/arocc for C processing now

21:20:16 Visen Yes but llvm is built into the compiler, so the experience is seamless. I don't need to install anything special to use the bindings generator. Also I think zig uses this https://github.com/Vexu/arocc for C processing now

21:20:31 shinmera anyway, it's a bit unfair to compare the two, tbh

21:21:07 shinmera it would be cool if bindings were less grunt work

21:21:13 shinmera but I also don't really mind so much? idk

21:21:35 shinmera I often find myself slightly changing naming or doing things slightly different to make it more amenable for lisp use

21:22:09 shinmera especially when it comes to gigantic APIs I pretty much never want to wrap all of it anyway

21:22:40 Colleen Yes hand coded bindings will be better, but autogenerated stuff has the advantage of being available instantly, so there is less friction when prototyping

21:22:40 Visen Yes hand coded bindings will be better, but autogenerated stuff has the advantage of being available instantly, so there is less friction when prototyping

21:22:56 shinmera sure

21:23:02 shinmera I'm just elaborating why I don't mind

21:25:37 Colleen shinmera: im mostly modifying code for testing and learning reasons, its how i figure out things. when i do fixes ill upstream of course

21:25:37 zyd shinmera: im mostly modifying code for testing and learning reasons, its how i figure out things. when i do fixes ill upstream of course

21:27:58 shinmera modifying the sacred script for devious experiments? how dare you

21:28:31 shinmera (but also you can still do that with ql code)

21:37:09 shinmera scymtym: I've added a feedback project for Trial with tracks for the engine/alloy/ide

21:37:13 shinmera https://feedback.tymoon.eu/trial/

21:37:22 shinmera should be a bit easier to keep track of stuff this way

21:44:19 Colleen I can do that ql code but then I will easily look track of changes, or forget etc. At least a magit buffer will easily show me "hey dumbass, remember that stupid thing you added? yup, still there."

21:44:19 zyd I can do that ql code but then I will easily look track of changes, or forget etc. At least a magit buffer will easily show me "hey dumbass, remember that stupid thing you added? yup, still there."

21:44:26 Colleen lose*

21:44:26 zyd lose*

21:46:00 Colleen I now have a beautiful 1.6GB directory of shirakumo-ware in git.

21:46:00 zyd I now have a beautiful 1.6GB directory of shirakumo-ware in git.

21:46:16 shinmera gonna guess most of that is the assets

21:46:27 Colleen 655M trial-assets

21:46:27 zyd 655M trial-assets

21:46:30 Colleen a big chunk

21:46:30 zyd a big chunk

21:47:48 Colleen Also kandria in there, which is 300MB

21:47:48 zyd Also kandria in there, which is 300MB

21:48:13 shinmera ah, so you're also counting the assets there

21:48:48 Colleen cl-gog-galaxy is 70M, surprising

21:48:48 zyd cl-gog-galaxy is 70M, surprising

21:48:54 shinmera it includes bins

21:49:02 shinmera many of the wrapper projects are huge because of that

21:49:08 Colleen makes sense

21:49:08 zyd makes sense

21:49:21 shinmera cl-mixed is 337MB :shepface:

21:49:53 Colleen but yeah, the actual size of just source code is much small but dunno how to count that. i just ran du -hs on my dir to get the size so not really accurate lol

21:49:53 zyd but yeah, the actual size of just source code is much small but dunno how to count that. i just ran du -hs on my dir to get the size so not really accurate lol

21:50:03 Colleen smaller*

21:50:03 zyd smaller*

21:50:14 shinmera find + du

21:50:59 Colleen cl-mixed is only 102M

21:50:59 zyd cl-mixed is only 102M

21:51:04 Colleen wonder why the size difference..

21:51:04 zyd wonder why the size difference..

21:51:21 Colleen fasl files?

21:51:21 zyd fasl files?

21:51:33 Colleen but that'd be an insane size blow up

21:51:33 zyd but that'd be an insane size blow up

21:52:10 shinmera zyd: history, prob

21:52:42 Colleen ah

21:52:42 zyd ah

21:53:32 shinmera (reduce #'+ (directory "~/Projects/cl/**/*.lisp") :key (lambda (p) (with-open-file (s p :element-type '(unsigned-byte 8)) (file-length s))))

21:53:49 shinmera god I need to add a file-size function to filesystem-utils that isn't as stupid as CL's

21:54:15 shinmera but, tomorrow

21:56:20 Colleen 11815162 smol

21:56:20 zyd 11815162 smol

21:58:44 shinmera 23122444 on my local checkout

21:59:11 shinmera though I only have like 19 projects checked out here

21:59:23 shinmera no, wait, that can't be right

21:59:36 shinmera 175, that's more like it

21:59:45 shinmera still, not all of em