libera/#clim - IRC Chatlog

7:06:01 GalaxyNova Does anyone know how to get rid of scroll bars from panes?

7:08:22 jackdaniel :scroll-bars nil

7:08:30 jackdaniel if you talk about the application pane / interactor pane and such

7:08:34 GalaxyNova thx

7:08:37 jackdaniel other panes don't have scrollbars by default

7:08:40 jackdaniel sure

11:53:02 cmack jackdaniel: I think I've determined the cursor position does update on clim-stream pane if it doesn't delegate to the text-input/preedit path... So I just need to determine how that's being subverted and emulate a similar update.

11:53:47 jackdaniel ah, OK; my hypothesis (as noted on codeberg) is that input editing stream has its own editing cursor

11:53:48 cmack sadly, I have no lisping time until a few hours from now

11:54:32 monash i have to figure out how to try the other non-x11 backends

11:54:33 cmack ah, I missed seeing that when I first checked earlier

11:54:40 jackdaniel I wrote it a moment ago

11:54:47 jackdaniel monash: (ql:quickload "mcclim-sdl2")

11:54:55 jackdaniel (setf clim:*default-server-path* '(:sdl2))

11:54:59 jackdaniel you should be set

11:56:52 jackdaniel I'm dividing my attention between finishing clipboard and being an unpleasent person on IRC who is skeptical of the brave new world of llm substituting for work (:

11:57:22 monash i haeve mcclim-sdl2/{slow,fast} (from a few days ago)

11:57:36 monash i'll try full

11:58:13 jackdaniel monash: note that some things may change with sdl2 backend (alpha warning is not for nothing :) but it works reasonably well. and with cmack improvements it will support keyboard layouts (finally!)

11:58:40 cmack I also tried using *original-stream* from the interactor-pane yesterday but no luck -- admittedly a shallow REPL experiment. But your comment is interesting idea

11:59:09 jackdaniel it is the other way around (unfortunately!) -- the original stream is the interactor pane

11:59:56 jackdaniel the input editing stream wraps around the original stream (encapsulates it)

12:00:27 jackdaniel or did I get that wrong?

12:00:27 jackdaniel hm

12:00:57 jackdaniel yes, you are right and I'm wrong, ignore me. original stream is bound to the encapsulating stream

12:01:07 monash i hit an assert on ccl when compiling: CODE = 127 CHAR = #\Backspace KEY-CODE = 8 KEY-NAME = :BACKSPACE (127 is "rubout")

12:05:51 jackdaniel I seem to recall this issue

12:06:01 jackdaniel yup, I've even commented on it!

12:06:02 jackdaniel ;;; XXX don't use the character #\delete because some (ekhm CCL) implementations

12:06:02 jackdaniel ;;; think that it is the same as #\backspace. To avoid confusion use #\rubout.

12:06:31 jackdaniel where does the assertion originate from?

12:08:45 jackdaniel does changing #\delete to #\rubout in Backends/SDL2/text-input.lisp help?

12:10:51 monash yes it did

12:12:10 monash i guess i already had the clx in this image

12:12:37 monash is there a top level way to shut a frame from another thread

12:13:16 monash just kill the thread i guess

12:15:20 jackdaniel in any case I've pushed to master the fix #\delete -> #\rubout

12:19:10 monash *all-ports* still only has the clx port

12:21:16 jackdaniel https://turtleware.eu/static/paste/ff7fb43b-foo.repl

12:21:37 jackdaniel *all-ports* contains all /running/ ports, not all /available/ ports

12:21:40 monash ok got it

12:26:38 monash esa menus trigger some bugs and i see my old friend "sdl thread not running"

12:28:56 monash i mean "sdl2 is not initialized"

12:31:11 monash one wrong restart ... and you're toast

12:34:49 cmack monash: but (destroy-port (find-port)) helps right? or do you feel like you have to restart your lisp image?

12:36:17 monash no it's workable without restarting the image, so it's cool

12:36:28 monash yes

12:37:03 monash (that and fudging mcclim-sdl2::*initialized-p* and calling %quit-port)