freenode/#clim - IRC Chatlog
Search
13:14:02
scymtym
setting PANE-BACKGROUND at a later point in time works, but using the :background initarg causes an error in REALIZE-MIRROR-AUX which seems to expect a color, not a design
13:19:13
scymtym
ha, instant scientific and sophisticated look: https://techfak.de/~jmoringe/protocol-analyzer.png
13:29:57
jackdaniel
that you can't pass design as :background? well, I'd certainly like to be able to pass it, so yes
13:48:31
jackdaniel
I think that this IF is bogus whatsoever: if ink can't be background, it should signal a condition like it does with permanent-medium-sheet-output-mixin
13:49:51
jackdaniel
(rgba-value ink) returns rgba color (I'm still not sure what format will be the best)
13:50:25
jackdaniel
for uniform inks it will return ink itself, but for patterns it returns matching design (etc)
13:51:13
jackdaniel
but it doesn't have to be a case (as you've noted with setfing explicitly pane-background afterwards)
13:51:27
scymtym
yes, it is only used for the background of the X window, though, which shouldn't matter too much
13:53:18
jackdaniel
but we'd have to check that (i.e for pane, which doesn't have always-repaint-background-mixin)
13:54:20
scymtym
i guess in the old times, or when forwarding X over a slow connection, it could have the benefit of drawing a correctly colored background onto your screen before drawing the window contents
13:55:08
jackdaniel
more I dig into clx more I'm convinced we should slowly migrate to xrender and abandon gcontext-based rendering
14:05:27
scymtym
jackdaniel: https://techfak.de/~jmoringe/x-background.png is with static cyan background for the X window
14:08:41
scymtym
there seems to be an issue with X coordinates as well: when the pane with the grid background grows large enough, an error "The value -32789 is not of type (SIGNED-BYTE 16) when binding XLIB::VALUE" is signaled
0:46:47
slyrus1
argh... after a few hours of trying to fix whatever was crashing my new clim demo/test program I think I once again have a new SBCL bug :(