freenode/#clim - IRC Chatlog
Search
10:09:50
ck_
loke: I'm not aware of any (but that doesn't mean much of course). Closest thing I can think of is the constant repainting of interactively drawn things, like in the paint demo
10:10:22
loke
ck_: Yeah, but that runs in a tight loop with a sleep. I need to be able to interact with the UI at the same time.
10:10:49
ck_
hehe, it doesn't work at all right now -- "Attempt to invert the probably singular transformation #<CLIM-INTERNALS::STANDARD-HAIRY-TRANSFORMATION 2 0 0 0 174 83>."
10:14:37
loke
In my case, I'm building some basic animation functionality into cliamxima. When I use it to solve the equiations of motion for a system using the lagrangian, I want to be able to use the result directly and animate what the motion looks like.
10:18:09
scymtym
loke: logic-cube has an example (don't copy that one). stopwatch has an example. RADAR-PANE in gadget-test is an example
10:19:10
scymtym
logic-cube is bad (among other things) because it blocks the event loop while playing the animation
10:36:27
scymtym
loke: around 4:00 this shows how the animated demos interact with "global" double buffering performed by the backend: https://techfak.de/~jmoringe/mcclim-gdi-backend-progress-1.ogv . for the demo, i disabled the "local" double buffering in logic-cube (there his been some more progress since i recorded that video)
10:38:40
loke
I just implemented animation by having a thread that loops, then it regularly calls EXECUTE-FRAME-COMMAND which updates the output records and calls REPAINT-SHEET. The call to EXECUTE-FRAME-COMMAND ensure that the update happens in the main thread.
10:40:31
scymtym
there is a mechanism for receiving timer events. if the thread does nothing else, you could use that instead
10:43:04
loke
scymtym: CLIM-INTERNALS :-) jackdaniel is sure going to love it when climaxima adds even more internal dependencies :-)
10:44:05
scymtym
hm, i just saw that as well. i guess we should either export it as an extension, remove it from the example or replace it with a full animation system
10:44:55
loke
scymtym: Have you had a look at the form input? It's pretty broken these days, which is a huge hassle for climaxima.
10:46:27
scymtym
no, sorry. i'm mostly working on backend and drawing things. but i'm also experiencing unresolved input-related issues
10:47:19
jackdaniel
accepting-values requires heavy refactor, some paving stones has been set for that
10:47:51
loke
Some time ago (a year?) the ability to go back in the form stopped working. If you have a command with two arguments, you can edit the first, then edit the second, but then you can never go back and edit the first one again.
10:48:47
jackdaniel
I think that saying "accepting-values operator is broken" is less ambigous than "input form"
10:48:57
loke
jackdaniel: Do you recall the discussion we had some time ago about the separator character used for inputting multiple arguments? Today, if you have a command that accepts more than one argument, you have a problem if you want to enter a space in the argument.
10:49:31
jackdaniel
it is specified in the spec, when you have spaces, you need to surround whole argument in ""
10:50:20
jackdaniel
write a proposal and we may discuss it; that would require adding new code, not simply ignoring some constraint
10:51:39
loke
Well, the thing is that current behaviour isn't compatible with the spec even. The spec suggests that space should move to the next argument, right? However, right now typing a space will insert a space into the argument and there isn't any way to move to the next.
10:51:40
jackdaniel
scymtym: if I understand correctly, loke argues that delimiter gestures should be ignored, but I don't get how that would work in broader picture
10:52:05
scymtym
so for a command that takes two required arguments, when i type the command name, type the first argument, type the second argument, should i be able to move the cursor to the first argument and edit it? or delete the second argument to get to the first?
10:52:08
loke
I'd like to suggest using TAB to move to the next argument. This would be consistent with modern UI's as well.
10:52:55
scymtym
or are we talking about the case when a translator for a two-argument command is invoked and an accepting values "mini dialog" is inserted into the interactor?