freenode/#clim - IRC Chatlog
Search
11:03:24
scymtym
dk_jackdaniel: thanks for looking at the branch. assuming everything works, should i clean the commits up in a new branch?
11:10:01
dk_jackdaniel
yes, please, then I'll look at it once again and after merging I'll start working on ieh-2
11:11:00
scymtym
sounds like a plan. one thing i'm not sure about are menus. they seem to work reasonably in the branch, but i'm not certain their behavior is completely unchanged
11:14:47
dk_jackdaniel
we are going to change their behavior to click-to-focus soon™ anyway, so given they seem to work I wouldn't stress too much about that
11:34:13
dk_jackdaniel
regarding postscript, I don't really like that it depends on clim-core, it seems like the order of dependencies is reversed (because postscript doesn't really require commands neither work with them)
11:35:14
dk_jackdaniel
ideally all backends would build on top of abstractions defined in clim-basic without any knowledge of what is in clim-core (although frame-managers would require extra thought with this regard)
11:37:32
scymtym
we could make the basic output destination mechanism independent of commands to achieve that, but ACCEPTing output destinations is arguably a part of the mechanism so that wouldn't work super well
11:38:58
dk_jackdaniel
maybe we could maintain a port registry with some characteristics and output-destination could use that. right now we do depend on symbol property lists, but having a real protocol to register backends would be more obvious solution
11:41:01
scymtym
i'm not prepared to tackle that problem right now. but let me suggest https://github.com/scymtym/architecture.service-provider for later
11:42:38
scymtym
i will probably finish my general testing pr first. the postscript/pdf/rasterimage one already has too much stuff in it
11:46:42
dk_jackdaniel
I even have a macro somewhere which creates a set of functions to manipulate a certain collection
16:34:27
scymtym
jackdaniel: i made https://github.com/McCLIM/McCLIM/pull/845 for the test stuff. i think it makes sense to work on this one first since it has the PRINT-TEST-PAGE utility which the other pr could use for the backend tests