tynet-lichat/shirakumo - IRC Chatlog
Search
2:59:02
Colleen
Shinmera: Here's my little review of my play through of Kandria https://dpaste.com/FVTTE8QPY
2:59:02
jmes
Shinmera: Here's my little review of my play through of Kandria https://dpaste.com/FVTTE8QPY
6:28:53
|3b|
as in a compound glyph which contains a compound glyph that contains the first compound glyph
6:29:42
|3b|
(or something like that, didn't investigate in detail, just that it was looping... put in a check for that so i could get back to testing sdf stuff and will investigate further later)
6:32:23
|3b|
ACTION found a pile of ttf files on another system which has a bunch of copies of old filesystems, so testing full sdf generation on ~1650 or so fonts
8:30:21
shinmera
This fella in the emails is kinda annoying saying there's bugs but not actually which
8:31:43
shinmera
he had written me an email before saying he'd been patching stuff (but not what or why)
8:33:57
shinmera
Like, they somehow thought sdf/bmfont was not lisp? and also there's bugs in 3d-v/m but they never reported them. https://plaster.tymoon.eu/view/3435#3435
8:34:28
shinmera
Yes, I saw kaveh in the lisp discord posting questions about his new gui lib and I wanted to tear my head off because we really don't need yet another fractured attempt
8:37:38
|3b|
ACTION apparently has ~2.6M distinct glyphs among the 3000 or so distinct truetype font files i have (at least the readable ones)
8:40:10
|3b|
probably a lot of nearly duplicate fonts in that set though, different versions of same fonts of whatever
8:40:50
shinmera
Good progress though, and good on you for trying to cover everything instead of releasing and getting reports from distraught users later (the "fuck around and find out" strategy that I employ)
8:42:29
|3b|
looks like a good # of my errors so far are glyphs with a single contour containing a single edge from 0,0 to 0,0 or similar :/
8:44:36
shinmera
I feel like this is one of those cases where having more validation on the creation side would have helped. It seems very analogous to the pains of writing an HTTP server and encountering all the wrong clients out there.
8:51:18
|3b|
looks like i found more bugs in the contour cleanup stuff too though, hopefully not too annoying to fix