freenode/#clim - IRC Chatlog
Search
7:31:23
jackdaniel
but today I'll focus on squashing GC clip issue and optimizing clipping to a polygon
7:32:38
jackdaniel
I think that they should be added lazily (i.e in slot-unbound method when accessed)
7:32:44
slyrus
OK, I think I agree, but I have more vague feelings than principled complaints at the moment.
7:33:30
slyrus
It's good that you're focusing on the GC clip and poly clipping, but I wouldn't necessarily let those hold up landing the xrender-migration branch! I've been using it and there's a lot of good stuff in there.
7:33:33
jackdaniel
it doesn't bother me much right now, but if you care about that think it over and create an issue so it won't get lost
7:34:02
slyrus
My other big annoyance, at the moment, is the flickery nature of the CLX backend, but that's, again, a problem for another day. Will definitely need to get solved though.
7:34:14
jackdaniel
I know there is a lot of good stuff, but I'm not very eager to merge branch which may land programmer in a debugger when they scroll the pane with polygonous clipping area
7:35:25
jackdaniel
preveiously we didn't create such clipping regions in any demo (and even if we did, we didn't scroll them)
7:36:14
jackdaniel
(nb: it is another defect I plan to squash: output records always have clipping region, even if the initial drawing operation didn't)
7:36:33
slyrus
Did you see my workaround for the other flicker issue? Where I make the panes up in the :panes section and then switch between layouts? Not optimal, but functional, at least, and allows for editing of various parameters of the shapes in my clim-paint app.
7:40:39
jackdaniel
please make an issue about panes, I belive this will be part of a task I've mentioned to scymtym yesterday (with frame managers)
7:41:01
slyrus
OK. Busy with real work at the moment, but I'll get to it in the next couple of days.
7:41:20
jackdaniel
I'll let myself to post offtopic link (sorry upfront!): https://blog.mozilla.org/blog/2018/12/06/goodbye-edge/
7:42:03
jackdaniel
how do you distinguish real work from imaginary work? ;-) I'm off too, squashing bugs
12:05:45
jackdaniel
(decode-type (or (member :none) pixmap) clip-mask) ; ← seemingly innocent, but apparently if clip-mask is not of type (member :none) we have an unmet assertion
12:07:20
jackdaniel
problem remained unnoticed for a big part, because setting pixmap works OK(!), debugger pops when we try to query the clip region from gcontext
12:10:30
jackdaniel
(down the rabbit hole problems is caused by a macro expansion which puts a wrong type declaration, so it may work on implementations which do not expand declarations into type checks)
13:52:46
dim
using http://godsnotwheregodsnot.blogspot.com/2012/09/color-distribution-methodology.html for now, a static curated list
13:53:13
jackdaniel
dim: I've read recently a little about colors. in principle we could come up with an algorithm which computes inks - more of them dimmer (no pun intended) they must be
13:54:14
jackdaniel
https://pavelkosenko.wordpress.com/2015/08/13/what-is-good-color/ ← I think someone from here linked this piece to me
13:54:25
jdz
Supposedly there's an ISO standard or something, and circe (the emacs IRC client I use) uses it to colour nicks.
14:08:03
dim
well anyway I found a list of colors, and could debug my program by generating another simple visualization using McCLIM
14:48:15
dim
I have way to much fun doing those quick 2-d hacks to represent a program's data structure I'm working on
14:49:28
beach
I think I just made a mistake. I was in the xrender-migration branch, and I did a git pull using the master branch. Now I have conflicts in Backends/RasterImage/package.lisp.
14:49:30
jackdaniel
I'm desecrating clx dependent module by removing bits for genera, lucid and texas instrumetns explorer
14:49:53
jackdaniel
beach: in that case: git reset --hard origin/xrender-migration (or git reset --hard origin/master)
14:55:38
jackdaniel
dim: regarding contarsting colors: https://randoma11y.com/#/?_k=fcf0ju this is fun to try
16:33:33
dim
the second one is calling into draw-rectangle hundreds of thousands of times, that's very slow, but I don't care very much and it gives a nice animation