18:08:08frgoQuestion to the CLX gurus here: In 'buffer-input (in file buffer.lisp in sharplisper's CLX), when a timeout occurs, the buffer is closed. I do not understand this - if is not an error, just a timeout ...
18:24:38jackdanielfrgo: try replacing allegro-specific buffer-input-wait-default with the one in dependent.lisp
18:25:17jackdanielI think that the order of cond clausesand actual wait implementation matters
18:25:41jackdanielas of why buffer is closed on timeout, I can only guess that it is by design: if X fails to respond in timely manner it is considered dead
19:00:38jackdanielafaik he is not interested in clim whatsoever
19:06:11ck_We'll see where this goes. Through the power of detachment, I can just sit back and watch
19:07:33jackdanieljudging what I've read here this fix makes sense
19:08:12jackdanielso I wouldn't be bothered much. that said we may indeed work on code on our side to call this function less if it is uncommon in x11 protocol `standards'
19:08:18scymtymthe comment sounded like clx might already have some kind of cache
19:08:28ck_I believe his point is that the caching already happens, and that querying the depth so often is a mistake
19:09:11ck_what I don't understand is why the difference is so great between linux and macos.
19:09:40scymtymthat too, but the server roundtrip in itself is an issue, on linux as well
19:14:34ck_I had a very quick look at the macro (attributes.lisp+309), but couldn't see where the depth is cached in those related functions
19:33:19ck_best get some sleep -- good night everyone