freenode/#clim - IRC Chatlog
Search
7:56:57
jackdaniel
thank you. it was indeed nice, however a little tiring. either way now we know, that the little one is very much capable of traveling to Zurich ;)
7:58:45
jackdaniel
yes, I've booked a hotel and plane tickets already (we arrive in the evening of 25th and leave in the morning of 30th)
7:59:44
jackdaniel
a pleasant bonus is that ELS will cover at least two days in the hotel and my part of the plane tickets
8:02:14
jackdaniel
scymtym: thank you for the review, I'll rebase the branch and go over issues one by one after that (and try some other demos)
9:55:08
jackdaniel
we've looked at the price and took "Swiss Star California" immedietely after that ,) which is 25m walk from the conference site which is nice
9:58:24
jackdaniel
huh, interesting, gc invariant fails only when sbcl is run from the console, when started from slime it works OK
9:59:25
jackdaniel
I'll need to check that on newest sbcl version to see if the issue is still present
10:29:17
beach
So I thought I had missed a closer hotel. Google maps claims that it's a 3 minute walk for us.
11:29:48
jackdaniel
Sounds good, however I'm little cautious of how patient the little one could be, so it might be that only I could arrive. We need to discuss it with Justyna.
15:00:37
jackdaniel
scymtym: regarding with-double-buffering, I remember you had your own implementation - how does it work?
15:01:24
jackdaniel
I'd like to get rid of double buffering as it is defined in graphics.lisp (it is a hack), that should be solved at medium's discretion
15:03:16
jackdaniel
n.b double buffering as it is defined now gets in a way with xrender extension and the ttf renderer
15:06:19
scymtym
jackdaniel: the removal in this pull request is due to a race condition when the mirror, medium and transformation is temporarily switched back and forth. my longer-term goal is completely different from the current implementation
15:08:00
jackdaniel
yes, race conditions with access are precisely the problem when used with xrender. I'm asking here, because question is not relevant to the pull request itself
15:09:41
scymtym
my "proper" implementation uses an offscreen buffer in "single-mirroring" mode for each top-level sheet. the port then draws complete frames at a given rate by collecting and drawing these offscreen buffers
15:11:19
scymtym
the complicated part is not collecting an offscreen buffer while it is in an inconsistent state. to ensure this, the port submits an event to the sheet in question which, when it is processed, blocks event processing for the sheet until the port is done
15:13:09
jackdaniel
n.b regarding meta-key / left-alt -- while we do not consider keyboard layotus now (there is are long-standing tickets on both mcclim and clx tracker about that), in principle in some languages (i.e Polish), alt-left + [character] produces i.e ś (or other language-dependent characters), and that behavior is very much unlike meta modifier. I wonder if the second commit won't have to be eventually
15:15:25
jackdaniel
yes, but we were discussing it from another angle (I think I didn't mention the keyboard layout thing)
16:00:54
jackdaniel
scymtym: I've reviewed 868 and left comments. while doing that I've found a bug (which is not a regression in 868) -- it is reported as 961