tynet-lichat/shirakumo - IRC Chatlog
Search
17:23:55
shinmera
Well, I think I got the bot working to the extent I need it to. https://github.com/Shinmera/stranger
17:24:09
shinmera
haven't done the actual import yet cause a buncha folks have not filled out their god darn surveys still.
17:29:40
shinmera
and -1000000% patience for things that haven't been updated in 3 years for a platform as shakey as discord
18:41:35
Colleen
drakonis: exactly. I was also curious about what define-page did, and I cant go to definition either.
18:41:35
Kingsy
drakonis: exactly. I was also curious about what define-page did, and I cant go to definition either.
18:42:03
Colleen
I feel like all that needs to happen is radiance needs to be quickloaded into the repl sly is using or something.
18:42:03
Kingsy
I feel like all that needs to happen is radiance needs to be quickloaded into the repl sly is using or something.
18:49:26
shinmera
I can assure you radiance does nothing special, so the lack of completions is a global lisp issue with your setup.
18:52:13
Colleen
https://github.com/doomemacs/doomemacs/blob/master/modules/lang/common-lisp/config.el#L45
18:52:13
drakonis
https://github.com/doomemacs/doomemacs/blob/master/modules/lang/common-lisp/config.el#L45
18:53:56
Colleen
hm so I just did M-x find-variable - flex, and I don't see that, how are you setting it?
18:53:56
Kingsy
hm so I just did M-x find-variable - flex, and I don't see that, how are you setting it?
19:04:52
Colleen
hmm I just will find it mad if its the completions. how has no-one had this issue before? Any package outside of builtins doesnt get gotos?
19:04:52
Kingsy
hmm I just will find it mad if its the completions. how has no-one had this issue before? Any package outside of builtins doesnt get gotos?
19:05:25
Colleen
https://github.com/doomemacs/doomemacs/blob/3766523f1ef0ae166bafe7c95be523d6304c1e5f/docs/getting_started.org
19:05:25
drakonis
https://github.com/doomemacs/doomemacs/blob/3766523f1ef0ae166bafe7c95be523d6304c1e5f/docs/getting_started.org
19:06:50
Colleen
https://github.com/doomemacs/doomemacs/blob/3766523f1ef0ae166bafe7c95be523d6304c1e5f/docs/getting_started.org#configel
19:06:50
drakonis
https://github.com/doomemacs/doomemacs/blob/3766523f1ef0ae166bafe7c95be523d6304c1e5f/docs/getting_started.org#configel
19:08:11
Colleen
Kingsy: (after! sly (setq sly-complete-symbol-function 'sly-flex-completions)) this works
19:08:11
drakonis
Kingsy: (after! sly (setq sly-complete-symbol-function 'sly-flex-completions)) this works
19:11:32
Colleen
right NOW I get completion on define-page but I still cant goto definition. but I think its because of the way I am calling the function
19:11:32
Kingsy
right NOW I get completion on define-page but I still cant goto definition. but I think its because of the way I am calling the function
19:12:39
Colleen
yeah it is, if I add radiane: before it, it works, why might that be? it must know because of the in-package and the asd ?
19:12:39
Kingsy
yeah it is, if I add radiane: before it, it works, why might that be? it must know because of the in-package and the asd ?
19:12:47
Colleen
you can also set (company +childframe) on your doom config to get a nice little box that gives you information about the completion
19:12:47
drakonis
you can also set (company +childframe) on your doom config to get a nice little box that gives you information about the completion
19:25:00
Colleen
yeah i did that, I just don't see any difference, hmm perhaps I just havent noticedwhat it does.
19:25:00
Kingsy
yeah i did that, I just don't see any difference, hmm perhaps I just havent noticedwhat it does.
19:53:50
Colleen
drakonis: any other cool quick wins in doom that will make my life easier when starting with lisp?
19:53:50
Kingsy
drakonis: any other cool quick wins in doom that will make my life easier when starting with lisp?
19:55:24
|3b|
not only is that random notch about 1/1000 of an em deep, the top edge goes left 1 unit, then back right 1 unit, then back left 1 unit
19:56:34
shinmera
is that just people having weird crap in their outlines that they couldn't see in the editor?
19:56:44
|3b|
which leaves that extra yellow line, which is a completely separate contour, containing only a diagonal line and a bezier that is the exact same as the line :/
19:57:35
|3b|
ACTION suspects some of the random stuff is people "signing" fonts (and/or marking them for copyright reasons)
19:58:22
shinmera
I know some use stuff like zero width spaces in text to detect if someone copypastes
19:59:04
|3b|
the brighter line with squares on it is a piece of a glyph, zoomed in by a few thousand
20:00:24
|3b|
for scale, the horizontal line by the cursor would be about 1mm if you printed the character 1m wide
20:02:38
|3b|
where it looks like something is very wrong with the code, but no it is just like that in the font
20:05:02
|3b|
ACTION wonders if i can easily distinguish odd #s of backtracks like that from even #s
20:06:18
|3b|
(while just walking the contour and maybe looking at a neighbor or 2, not building some full index of all edges and checking for duplicates)
20:06:52
shinmera
that hecked up font reminded me of this https://www.youtube.com/watch?v=HLRdruqQfRk
20:33:48
shinmera
by deleting edges do you mean merging edges that are extremely close and have small control points?
20:43:58
|3b|
ok, handles that glyph correctly now i think? 1 edge left from the 3x duplicate one, and entirely removes the contour that was just a line and linear quadratic bezier version of the line
20:47:31
|3b|
and if that works, i can start trying to figure out where to hook in the part about actually returning cleaned up versions of the glyphs
20:48:53
shinmera
I know this is way too early, but any funny ideas about incremental atlas generation?
20:49:23
|3b|
hmm, left in a debug print about that case of degenerate contours, and that font prints it out a lot :/
20:51:21
shinmera
also a question of my renderer being smart enough and learning things like texture switching so it doesn't have to be a single texture.
20:53:19
|3b|
ACTION wonders how incremental SDF generation would look (generate 1/4 res, then fill in missing parts to make 1/2, then repeat again to get full res, or similar)
20:55:15
|3b|
aside from optimization (and possibly using some trivial binpacker like just a grid) that's about the only thing i can think of for runtime generation
20:56:13
shinmera
I can think of pre-allocating glyph space in the atlas and already starting text rendering while the atlas is being updated and uploaded in a separate thread.
20:57:17
|3b|
yeah, packing isn't quite as fast, but still probably good enough for that depending on how many you put into a single atlas
20:58:03
|3b|
(packing things in random order is pretty bad for efficiency anyway, so might not matter much for fonts)
20:59:21
|3b|
and doesn't really care if you give it 1 at a time or all at once. only thing that affects is that if you sort first you tend to get better packing, which you obviously can't do with 1 at a timer
21:03:03
shinmera
well, I look forward to having nice, fast rendering in Alloy in ten years from now when I'll probably finally get to doing all the bits for it :v
21:03:41
|3b|
but probably could just pack into a grid for a lot of text use and not lose too much space compared to unsorted "good" packing