libera/#clim - IRC Chatlog
Search
6:41:27
beach
contrapunctus: I added month names to Acclimation in case you want different languages.
8:04:43
contrapunctus
beach: thanks, didn't know about Acclimation. (Clever name 😄️) The other project I just got to know of recently is CLIMatis...it looks interesting, but it seems it hasn't been updated in a few years - what's its status?
8:05:40
beach
I wrote it when McCLIM did not seem to make progress, so I imagined a CLIM3 specification with CLIMatis being an implementation of it.
8:11:34
beach
CLIM3 had one interesting feature though, namely that it acted upon a remark in the CLIM II specification that suggested the unification of sheets and output records. But then I inadvertently introduced a limitation that turned out to be unacceptable, namely that instances of this unified concept had to have a rectangular shape.
10:35:05
loke[m]
jackdaniel: Do you mind changing CLIMI::+FONT-SIZES+ to a variable instead of a constant? I used to modify this one to have McCLIM properly adjust to 4k displays.
10:36:16
loke[m]
This is the code I used to run (which doesn't work anymore due to changes in SBCL which no longer allows changing constants)
14:23:04
jackdaniel
loke[m]: I don't really mind (it is an internal symbol, so we may fiddle with it as much as we want without guaranteeing anything) although font scaling should be probably handled by the backend?
14:26:19
jackdaniel
jolby: I'll look into your repository when I'm back home; as of me - no I don't have anything public re sdl backend except for what I've send you
15:21:15
loke[m]
<jackdaniel> "loke: I don't really mind (it is..." <- I agree it should be handled at the backend, and my proposal is that the function I shared should actually be part of the CLX backend, as it performs the appropriate adjustment of fint sizes based on the dpi properties.
15:21:42
loke[m]
I certainly would much rather add this to the CLX backend rather than having my code mess with the internals. Would you be OK with this?
15:23:35
loke[m]
A better solution may be to have the DPI property affect the font sizes in general (so that font size 10 is adjusted by the dpi setting at the time of assignment). It's been a while since I tried that, but I think there was a problem where some code assumed that a font size of N meant that the text was N pixels high, which is obviously not true anymore. Adjusting the standard font sizes turned out to work quite well in practice.
15:24:05
loke[m]
However, I'll be happy to try the more "correct" solution again to see what the problem with it was. Maybe I did something wrong last time.