tynet-lichat/shirakumo - IRC Chatlog
Search
11:06:42
shinmera
and an untested system for capturing the event state, and thus being able to generate "replays"
11:20:40
shinmera
Not all of it, mind you. Just whatever subsystems I feel are mature enough to document.
11:56:15
shinmera
scymtym: Also I'm considering taking time after november to localise Kandria into German myself.
11:57:07
shinmera
though I'm not sure yet if I would, since it *would* be a lot of tedious effort :v
12:04:40
Colleen
i'm asking because i think SBCL's ctors don't apply to REINITIALIZE-INSTANCE (claps has that optimization if i recall correctly)
12:04:40
scymtym
i'm asking because i think SBCL's ctors don't apply to REINITIALIZE-INSTANCE (claps has that optimization if i recall correctly)
12:06:13
shinmera
but yes, I'm not at all sure if it actually elides allocations like this. Might have to change it to just hammering slot-values instead.
12:07:55
Colleen
my mistake, i should have said INITIALIZE-INSTANCE. but the comment applies regardless
12:07:55
scymtym
my mistake, i should have said INITIALIZE-INSTANCE. but the comment applies regardless
12:12:51
Colleen
ACTION is interested in key to text translation since he recently implemented a big chunk of xkbcommon
12:12:52
scymtym
ACTION is interested in key to text translation since he recently implemented a big chunk of xkbcommon
12:13:34
shinmera
yeah it's a mess, but something the OS has to give you, since only it knows the keyboard layout in use and how it handles combinations, etc.
12:14:15
shinmera
there's also local-key-string callback function on the context for similar reason.
12:53:53
shinmera
Probably gonna merge the PR later today, since I *think* one of the bugs in Kandria's inputs is related to the event handling.
12:54:16
shinmera
as in, I'm fearing it dropping a key release event, leading the input to get stuck indefinitely.
15:40:58
shinmera
selwyn: scymtym: though, any opinions on putting the queue into luckless instead of Trial?
15:46:18
selwyn
i feel that data structures are potentially useful enough to others to be made available independently
15:50:57
shinmera
Ok, I'll do it like this: copy into luckless for now, and once it's in ql, switch trial over.
15:57:06
Colleen
Shinmera: dayjob and scope creep :) i had to change more things than i thought. i guess i can submit a draft so you can get an idean
15:57:06
scymtym
Shinmera: dayjob and scope creep :) i had to change more things than i thought. i guess i can submit a draft so you can get an idean
16:28:55
shinmera
scymtym: couldn't BINARY-TYPE-ENCODER be fixed more simply by using (ub8 '(lambda (v i e) (setf (aref v i) e))) instead of rewriting them all to lambdas?
16:30:55
shinmera
I would, it's a simpler change than the other one and shouldn't impact performance any.
16:32:08
shinmera
Other than that the changes look alright to me. Though I think the min-version thing ideally would be done more smartly, by checking what features are actually used (compression scheme, encryption scheme, zip64-ness)
16:33:00
Colleen
right, figuring that out is one of the reasons this took longer than i initially thought
16:33:00
scymtym
right, figuring that out is one of the reasons this took longer than i initially thought
16:37:42
Colleen
yeah. would you settle for logic of the form "ZIP64 used" -> version needed = a.b, "ZIP64 not used" -> version needed = c.d? because i don't think i have time to go over other features and find the corresponding versions
16:37:42
scymtym
yeah. would you settle for logic of the form "ZIP64 used" -> version needed = a.b, "ZIP64 not used" -> version needed = c.d? because i don't think i have time to go over other features and find the corresponding versions
16:39:11
shinmera
Maybe just a function "determine-min-version" and then do that in it for now with a FIXME note.
17:56:41
shinmera
Well, wrote cursory docs for luckless and submitted for ql. https://shinmera.github.io/luckless/
18:02:37
shinmera
I would really like it to have sequences support since the other functions are kinda annoying, but oh well.