libera/#clim - IRC Chatlog
Search
14:32:48
scymtym_
can anyone try what happens when evaluating (restart-bind ((crash (lambda () #P"foo"))) (error "bar")) and selecting the CRASH restart in the SLIME debugger?
14:41:32
scymtym_
it seems SLIME-EVAL just tries to read the result produced on the common lisp side as an emacs lisp expression :(
15:00:22
jackdaniel
http://turtleware.eu/static/paste/incrm.mp4 now this video exhibits behaviors with and without disarming the gadget
15:00:41
jackdaniel
if it is disarmed then the focus is immedietely lost, when it is not disarmed, then we have a flickerfest :)
15:20:13
jackdaniel
(defmethod note-sheet-disowned ((gadget basic-gadget)) (deactivate-gadget gadget))
15:49:07
jackdaniel
ind that wrapping it in updating-output may cause it to capture some "old" variable if it is shared with other records, that depends when and from where they are redisplayed
15:57:05
jackdaniel
yes, it is expected behavior. the old reference may be captured in a callback that is stored in the record, and the record does not change so the callback does not change either
17:36:06
jackdaniel
slyrus: I'm around for another hour, so if you want to go through some issues then we can go now
17:37:07
slyrus
one minor-ish issue that I've run across in the clim-listener is described here: https://github.com/slyrus/clorg/issues/1
17:37:50
slyrus
my larger issues resolve around how to get decent zooming/scrolling - and one day perhaps - more "interesting" drawing behavior based on the scrolling/zoom settings.
17:38:16
slyrus
one approach to scrolling/zooming and the immediate problem I run into is described here: https://github.com/slyrus/zoom-viewer/issues/1
17:40:22
slyrus
if I manage the scrollbars myself, rather than using a scroller-pane, I get "better" results: https://github.com/slyrus/zoom-viewer-2
18:07:53
jackdaniel
you may test this code. notice that I'm not changing the sheet-native transformation only the sheet-transformation
18:34:18
slyrus
ok, I'm reluctant to remove to much as it is likely in the present methods where McCLIM or I go wrong, but a slimmed down version of clorg that doesn't have any external dependencies can be found here: https://github.com/slyrus/standalone-clorg
18:34:56
jackdaniel
OK, and here is zoom with working scrollbars: https://plaster.tymoon.eu/view/2733#2733
18:34:57
slyrus
as for the zoom-frame, ok, I'll try to remove my use of sheet-native-transformation.
18:36:08
jackdaniel
it resets requirements without accounting for the transformation, so don't resize the frame - otherwise scroll bars will be re-set
18:37:17
slyrus
hmm.... it sort of works with scroll bars but changing the zoom still resets the zoom
18:49:34
slyrus
if I zoom in, then, say, scroll down so (that only) the middle portion of the zoomable area is shown, if I then adjust the zoom, the displayed area resets to 0,0 at the upper left
19:00:52
jackdaniel
this tangentially resembles this: https://github.com/McCLIM/McCLIM/issues/658 but I don't think that the underlying issue is the same
19:03:26
jackdaniel
I've got to go now, but I have something to work with - I can reproduce both issues; I'll let you know if I learn something
19:04:36
slyrus
and I've managed to remove my calls to internal climi routines with no loss of functionality, so that's good!
19:05:26
jackdaniel
keep in mind that resizing the frame will invoke the layout protocol and it doesn't take into account the sheet transformation, so scroll bars will get wrong again
19:06:31
jackdaniel
and here is a fix for mouse scroll in scroll bars: https://plaster.tymoon.eu/view/2734#2734
22:36:39
jackdaniel
if anyone has time and energy to comment on this analysis (most notably the section 'propagating the output changes') then I'll appreciate that a lot. this section applies to the part of the specification that is not implemented at all in McCLIM yet
22:37:37
jackdaniel
there is no rush with it, I probably won't be working much on the incremental redisplay this week