freenode/#clim - IRC Chatlog
Search
13:32:50
loke
jackdaniel: in your implementation of request-selection, what if there is no active sleection?
13:57:21
jackdaniel
this generalisation would work for arbitrary selections with optional backend specialisation
14:18:08
loke
jackdaniel: The X11 selection machnism is generalised in terms of clipboard/selection, etc. In fact, it support an infite number of "boxes". However, only selection and clipboard is typically used. The old concept of numbered "cut buffers" isn't supported by anything these days.
14:18:29
loke
And othe rplatforms are even more limited. Windows only supports clipboard, afaik. OSX has clipboard+selection
14:19:11
loke
I think having CLIM settle on clipboard and selection is probably enough. If we try to generalise further we may end up in a situation where the mapping from CLIM concepts to native OS concepts are unnatural.
14:20:23
loke
By the way, I managed to override the history stuff in interactor, and replaced it with my own implementation that doesn't interfere with the existing functionality. It's ugly though, and I reference a lot of non-exported symbols to make it work.
14:20:29
loke
Here's the code: https://github.com/lokedhs/maxima-client/blob/master/src/maxima-history.lisp#L52
15:49:24
jackdaniel
loke: having general mechanism in clim will make it possible to map onto underlying display server what fits. for instance for windows we map :clipboard (eql specialisation), for X11 and OSX we map :clipboard and :selection
15:51:02
jackdaniel
so it works as well as it does with hardcoded clipboard and selection but with less verbs to advertise
15:51:32
loke
jackdaniel: I'm not against the general idea. My very first protoype (still in a branch somewhere) used it. In my rewrite, I decided to make it simpler and simply lock down two "boxes"
15:52:15
loke
I did that, because I didn't want to bother with the added complexity, but other than that I have no particular preference.
16:12:43
jackdaniel
adding sanity check for valid coordinates revealed a hidden defect in stream printing interface (found by Nisar), that's good – that gives me idea how we can speed up the interface substantionally