freenode/#clim - IRC Chatlog
Search
22:18:46
jackdaniel
I have a basic demonstration of immediate mode gui implemented with clim-basic primitives, I think I'll write a blog post about it when it grows to something useful
22:22:15
jackdaniel
(only button interaction is implemented, also plenty of priting to the repl which starts the application)
4:36:10
loke
jackdaniel: I spent some time studiying the path library yesterday. It provides most of what we need up to a point: It provides an outline (or concave polygon) representing the shape (including lines, circles, whatever) that is needed. Thus, I should be able to use this and then implement a standard algorithm for turning a concave polygon to a set of triangles that can be rendered by, well Xrender or anything else that may need it.
4:37:01
loke
It's my intention to write a separate package that encapsulates this, so it can be used by Xrender, and later, by the SDL or OpenGL backends.
4:59:58
ck_
loke: tell me if you can use any help with that. I can say without much ego that computational geometry is on my list of things I know slightly more than nothing about.
5:23:35
loke
I experimented with this: https://gist.github.com/lokedhs/4afe548554083ee2c39684b1387858a0
5:24:58
loke
ck_: basically, the PATH library gives you the ability to construct pats with joins and caps that are compatible with what CLIM needs. When iterating over the PATH-ITERATOR, you get line segments and (optionally) arcs that describes the enclosing object.
6:05:03
loke
jackdaniel: The emphasis was more on the fact that I want to make it non-backend-specific
6:07:11
loke
I currently don't know what kind of algorithm to use to split the polygon into triangles. I'm just assuming there are existing optimal ones
6:09:08
loke
I have the polygon stuff doen already using Xrender. However, it currently only supports colours (no patterns, etc)
6:09:13
jackdaniel
loke: right, that's what climb package is for: mixins and utilities which may be shared by backend developers
6:10:31
loke
beach & jackdaniel: a pity the resolution of the recorded stream is low so that reading the text is hard.
6:11:36
loke
But I'm assuming it must have been better, or they would have mentioned it on the stream
6:13:24
loke
Hangout have something called “Hangouts on Air” that allows you to stream a meeting at full quality.
6:14:04
jackdaniel
then someone was recording their own hangouts screen and live-streaming it from their desktop
6:14:06
scymtym
i put the X resource leak fixes i found while playing with double buffering into a tiny pull request: https://github.com/mcclim/McCLIM/pull/796
6:15:12
jackdaniel
I had to use different computer, because google refused to let me on the conference without being logged into their product
6:15:56
loke
jackdaniel: Yeah, you mentioned it on the stream. I was like... WTF!? It's almost like they care more about people getting google accounts and handing over all their information than actually providing a useful service. Funny that.
6:17:22
loke
Did you hear about their new “Shoestring” (“shoelace”?) service they announced recently? It's an implementation of a corner of G+, only available in one city in the US.
6:19:06
jackdaniel
I don't personally care that much about google, not using it except some occasions other people are in their walled garden
6:19:38
loke
Not onyl does it implement the meetings feature of G+. It also seems to be a reimplementation of Google Schemer? (also killed, of course)
6:19:58
jackdaniel
(just like I did not use Windows except few occasions when I had to start a particular program and it didn't run on wine0
6:30:01
jackdaniel
mostly stylistical requests, but I have also one question about presentations, it seems to me that you wrap a presentation inside a presentation during redo
6:32:26
ck_
if that's the case, it's an oversight; it was that way before even during creation. I probably didn't fix it at all points in the code
6:40:47
ck_
I was wondering about line width style guides -- the linebreaks are probably there to not cross the 80 column line.
6:42:14
jackdaniel
it is not a rule set in stone, I'm trying to have average size 80, but if sticking to it uglifies the code then I make an exception
6:55:39
ck_
your comment on using case for the bezier control points won't work because the things aren't eql, I guess