tynet-lichat/shirakumo - IRC Chatlog
Search
Thursday, 10th of June 2021, 15:02:43 UTC
20:21:49
Colleen
Where is the sbcl config .el file in portacle to set the heap size ?
20:21:49
seok
Where is the sbcl config .el file in portacle to set the heap size ?
20:22:51
shinmera
all/emacsd/portacle/portacle-slime.el
20:25:49
Colleen
Can I just add (setq slime-lisp-implementations
20:25:49
seok
Can I just add (setq slime-lisp-implementations
20:25:50
Colleen
'((sbcl ("sbcl" "--dynamic-space-size" "1024")))) at the end ?
20:25:50
seok
'((sbcl ("sbcl" "--dynamic-space-size" "1024")))) at the end ?
20:31:56
Colleen
Heap exhausted (no more space for allocation).
20:31:56
seok
Heap exhausted (no more space for allocation).
20:31:57
Colleen
31981568 bytes available, 37235872 requested.
20:31:57
seok
31981568 bytes available, 37235872 requested.
20:32:21
Colleen
Thats only 40mb right ?
20:32:57
shinmera
That's what's left, not the total heap size.
20:33:19
Colleen
I added '((sbcl ("sbcl" "--dynamic-space-size" "4096")))) though
20:33:19
seok
I added '((sbcl ("sbcl" "--dynamic-space-size" "4096")))) though
20:33:41
Colleen
How do I check if this line is working ?
20:33:41
seok
How do I check if this line is working ?
20:37:11
shinmera
(sb-ext:dynamic-space-size)
20:38:59
shinmera
dunno how old/recent your portacle is, there was a patch to fix an issue with passing arguments to sbcl from the wrapper.
20:39:10
shinmera
I think that made it into the last relesae.
20:39:18
shinmera
don't remember for sure.
20:40:44
Colleen
yeah, I restarted portacle and dynamic-space-size isn't changing. I'll try updating portacle
20:40:45
seok
yeah, I restarted portacle and dynamic-space-size isn't changing. I'll try updating portacle
21:00:56
shinmera
not sure, then. only thing I can think of is to compile the launcher fresh.
21:01:43
shinmera
hm, no, the bug was fixed in 1.3b according to release notes
21:01:48
shinmera
so 1.4 definitely includes it.
21:02:24
Colleen
I dunno, I've been stuck trying to increase sbcl heap-size on portacle like forever : (
21:02:24
seok
I dunno, I've been stuck trying to increase sbcl heap-size on portacle like forever : (
21:02:43
shinmera
Sorry, not sure what's happening. I do distinctly remember fixing it...
21:02:52
shinmera
try running it directly from the command line
21:03:12
shinmera
that's fine, open a cmd navigate to the portacle root and try
21:03:32
shinmera
win/bin/sbcl.exe --eval '(print :hi)'
21:03:38
shinmera
just to try command options at all
21:04:14
Colleen
hm. does anything in portacle stay running after I close emacs ?
21:04:14
seok
hm. does anything in portacle stay running after I close emacs ?
21:04:27
Colleen
I noticed I can't rename the folder
21:04:28
seok
I noticed I can't rename the folder
21:04:28
shinmera
win/bin/sbcl.exe --dynamic-space-size 8192 --eval '(print (sb-ext:dynamic-space-size))'
21:04:49
shinmera
It registers font files with windows to ensure emacs has one to display
21:05:06
shinmera
unfortunately they don't deregister after closing and there's no good way to do it other than restarting (thanks windows)
21:05:33
shinmera
hmm, let me check something
21:06:47
shinmera
Ok, no, never added a 'fontunreg' utility.
21:08:34
Colleen
hm, keep getting end of file error
21:08:34
seok
hm, keep getting end of file error
21:08:59
shinmera
that might be cmd's fucked up argument parsing
21:10:33
shinmera
ok, so at least that part is fine, phew :)
21:11:05
Colleen
yeah, windows doesn't believe in single-quoting in most places
21:11:05
SAL9000
yeah, windows doesn't believe in single-quoting in most places
21:11:17
shinmera
selwyn: try M-x portacle-recompile
21:11:41
shinmera
I remember that it file-compiles the emacsd things and won't pick up changes in the source automatically.
21:12:14
Colleen
and remember that windows, unlike linux, usually delegates parsing of command line arguments to the application, rather than doing it in the shell. as is usual for windows, there are skeletons in the closets and exceptions to every rule...
21:12:14
SAL9000
and remember that windows, unlike linux, usually delegates parsing of command line arguments to the application, rather than doing it in the shell. as is usual for windows, there are skeletons in the closets and exceptions to every rule...
21:12:32
shinmera
Oh, wait, it's not interactive, so M-: (portacle-recompile)
21:13:08
Colleen
portacle-slime.el:10:7:Warning: ‘slime-complete-symbol-function’ is an
21:13:08
seok
portacle-slime.el:10:7:Warning: ‘slime-complete-symbol-function’ is an
21:13:09
Colleen
obsolete variable (as of 2015-10-18); use
21:13:09
seok
obsolete variable (as of 2015-10-18); use
21:13:09
Colleen
‘slime-completion-at-point-functions’ instead.
21:13:09
seok
‘slime-completion-at-point-functions’ instead.
21:13:21
shinmera
SAL9000: argparsing is uniquely fucked up in cmd though. Like holy hell do I have nightmares from before I did the C wrapper, trying to do shit with batch files.
21:13:28
shinmera
yeah that's whatever, ignore it.
21:14:21
Colleen
my heap size is a big heap now
21:14:21
seok
my heap size is a big heap now
21:14:41
shinmera
another ticket fixed at 23:00
21:14:48
Colleen
Now I can lisp with 8GB
21:15:07
Colleen
I'm surprised you are almost always available
21:15:07
seok
I'm surprised you are almost always available
21:15:11
shinmera
note that setting the heap size huge doesn't really matter. It's only pages allocated, won't actually eat that much ram residential.
21:15:22
shinmera
you could set it to 64GB or whatever.
21:15:39
shinmera
What can I say, I live at the computer.
21:16:15
Colleen
meanwhile trying to drive JACK2 from Python. I love how the API demands that ports are registered before the main loop, but other clients' registrations show up in the opposite order (first the client appears, then its ports)
21:16:15
SAL9000
meanwhile trying to drive JACK2 from Python. I love how the API demands that ports are registered before the main loop, but other clients' registrations show up in the opposite order (first the client appears, then its ports)
21:17:05
Colleen
not to mention that all callbacks (RT or otherwise) are called in a non-main thread, and it does the Qt-style thing of "you shall only call me from main thread"
21:17:05
SAL9000
not to mention that all callbacks (RT or otherwise) are called in a non-main thread, and it does the Qt-style thing of "you shall only call me from main thread"
21:17:25
shinmera
I love how jack automatically connects some fucking random ass ports whenever I fire up my carla plugin rack so I have to go back into claudia and manually rewire everything
21:17:27
selwyn
thank you for making portacle shinmera
21:17:33
selwyn
i would hate to do windows stuff without it
21:17:46
shinmera
I haven't had time to update it in forever, and nobody else seems to either :(
21:18:27
Colleen
I've somewhat given up on making MinGW be sane and am now using WSL2 with an X server
21:18:27
SAL9000
I've somewhat given up on making MinGW be sane and am now using WSL2 with an X server
21:18:42
shinmera
What's wrong with msys2?
21:19:00
Colleen
the emacs port has some weirdo corner cases, particularly in regards to TRAMP
21:19:00
SAL9000
the emacs port has some weirdo corner cases, particularly in regards to TRAMP
21:19:20
Colleen
I might have to re-test, though, because ssh: vs sshx:
21:19:20
SAL9000
I might have to re-test, though, because ssh: vs sshx:
21:19:25
shinmera
haven't had any bug reports about that in portacle at least :v
21:19:41
selwyn
sal9000: i seem to recall switching from one to the other solved most of my problems
21:19:42
Colleen
windows + lisp = bug : (
21:19:56
Colleen
seok: in general, or a specific one?
21:19:56
SAL9000
seok: in general, or a specific one?
21:20:01
shinmera
the lichat bridge is cool and all but it does suck that nick completion doesn't work
21:20:19
Colleen
Shinmera: as in completing bridged nicks?
21:20:19
SAL9000
Shinmera: as in completing bridged nicks?
21:20:20
shinmera
sal get yer ass over here already >:O
21:20:51
Colleen
as good a time as any to yank weelichat into production :)
21:20:51
SAL9000
as good a time as any to yank weelichat into production :)
21:20:51
shinmera
technically the bridge (as specified) does allow bridging joins/parts, but I avoided doing that on purpose in the current impl.
21:20:58
Colleen
here's hoping it doesn't knock the applecart over
21:20:58
SAL9000
here's hoping it doesn't knock the applecart over
21:20:59
Colleen
I've not gotten around to getting ssl or sqlite to work on windows sbcl
21:20:59
seok
I've not gotten around to getting ssl or sqlite to work on windows sbcl
21:21:07
Colleen
and winhttp is slow ...
21:21:31
shinmera
never had ssl problems just using a downloadable dll from one of the openssl links.
21:21:56
Colleen
I can't connect to wss:// on windows lisp
21:21:56
seok
I can't connect to wss:// on windows lisp
21:23:54
Colleen
Shinmera: remind me, what was the story with weelichat and ssl? did we get it working?
21:23:54
SAL9000
Shinmera: remind me, what was the story with weelichat and ssl? did we get it working?
21:24:23
shinmera
didn't have the time to test it enough to figure out whether the problem was on the client side, the server side, or both.
21:24:47
shinmera
the pieces are there though.
21:25:05
Colleen
what was the ssl port? I'm going to try it
21:25:05
SAL9000
what was the ssl port? I'm going to try it
21:25:18
shinmera
on the official server? Oh, that doesn't work at all.
21:25:40
Colleen
Welp. Who needs security anyway? Guess my tymoon.eu password is going out in the clear.
21:25:40
SAL9000
Welp. Who needs security anyway? Guess my tymoon.eu password is going out in the clear.
21:26:23
shinmera
the default server is set up to serve ssl on the default 1112 port if you want to try it locally. https://github.com/Shirakumo/ex-lichat/blob/master/config/config.exs
21:26:59
shinmera
there's instructions on how to set it up in the readme too.
21:28:39
shinmera
I have also been thinking about an extension to allow sending hashed/SASL passwords.
21:29:17
shinmera
though actually the handshake is not set up to allow such an extension, since extension exchange and password are done in the same update... hrm.
21:30:08
Colleen
maybe CertFP would be easier?
21:30:08
SAL9000
maybe CertFP would be easier?
21:30:14
Colleen
(i.e. recognise hash of client cert)
21:30:14
SAL9000
(i.e. recognise hash of client cert)
21:33:46
Colleen
23:32 =!= TyNET: The update was malformed and could not be parsed.
21:33:46
SAL9000
23:32 =!= TyNET: The update was malformed and could not be parsed.
21:33:48
Colleen
23:32 =!= TyNET: You do not have sufficient permissions to perform this action.
21:33:48
SAL9000
23:32 =!= TyNET: You do not have sufficient permissions to perform this action.
21:33:50
Colleen
23:33 =!= TyNET: The requested channel does not exist.
21:33:50
SAL9000
23:33 =!= TyNET: The requested channel does not exist.
21:36:26
shinmera
yeah, but it works fine :v
21:37:07
Colleen
I see #shirakumo as a blank buffer
21:37:07
SAL9000
I see #shirakumo as a blank buffer
21:37:40
SAL9000
aha, shirakumo vs #shirakumo
21:37:42
shinmera
yeah, no # in front on lichat
21:37:48
SAL9000
the irc reflexes are strong
21:38:08
shinmera
you probably also want to be in the lichatters channel
21:38:22
shinmera
(autojoin list is separated by double spaces)
21:38:35
SAL9000
guess autojoin didn't do its job properly; it was set to "lichatters" by default
21:40:13
Colleen
damnit, it still pings me despite the rewrite rule
21:40:13
SAL9000
damnit, it still pings me despite the rewrite rule
21:40:55
shinmera
I just set notifications off on the #shirakumo channel
21:41:15
Colleen
iirc /set weechat.notify.irc.libera.#shirakumo none
21:41:15
shinmera
iirc /set weechat.notify.irc.libera.#shirakumo none
21:41:32
shinmera
since everything's here anyway, no need for notifications on the irc relay.
21:41:55
shinmera
actually, on that note, not sure if we need to do anything to respect those settings? (probably, right?)
21:43:28
SAL9000
we *should* be OK on that front
21:44:28
SAL9000
there we go. a different trigger, one that hacks up the raw irc message, keeps it from pinging me :)
21:44:41
SAL9000
nope, that kind of self-ping still works
21:45:04
SAL9000
//trigger add colleen2 modifier "irc_in2_privmsg" "${nick} == Colleen" "/^(:)(Colleen)(!\S+@\S+ )(PRIVMSG #shirakumo :)<(\S+)> (.*)/${re:1}${re:5}${re:3}${re:4}${re:6}/" ""
21:45:36
SAL9000
interesting, looks like we need to handle the // trick ourselves... I thought that was a builtin. Guess not.
21:45:55
shinmera
why is so much stuff just hard-coded
21:46:07
SAL9000
well it is an IRC client first-and-foremost, I guess.
21:46:08
shinmera
especially the smart filters still piss me off
21:46:24
SAL9000
the bit where we're expected to insert the tags for that ourselves?
21:46:56
SAL9000
right. I can see it being useful for sufficiently weird protocols with their own definition of "smart" but gaaah.
21:48:23
shinmera
then the filter can just be disabled for those channels.
21:48:53
shinmera
Having SSL working would be aces
21:49:16
shinmera
I think I have it disabled on the live server outright because it had issues grabbing the cert files and I didn't wanna bother.
21:49:28
shinmera
gotta get the local version working first anyway.
21:49:30
selwyn
did you ever find out why the old kandria discord got banned
21:49:37
SAL9000
does /lichat user-info and /lichat channels work for you?
21:49:44
SAL9000
the former seems to be a no-op and the latter trips "malformed update"
21:49:57
shinmera
the discord didn't, I did. And because I was the only admin, it got ghosted. And no, they never got back to me, of course.
21:50:16
shinmera
SAL9000: don't think so. I remember those being broken
21:50:44
shinmera
there's a couple commands that I never tested through.
21:51:52
shinmera
i know send, react, and edit work. I tested and fixed those recently
21:52:30
SAL9000
how do you do multiline?
21:52:45
shinmera
what, if it's longer than it used to be?
21:52:56
SAL9000
I mean \n inside messages
21:53:26
SAL9000
\o/ not a big deal. I just noticed that we're setting input_multiline...
21:53:47
shinmera
(defmethod register-object-for-pass (pass (editor editor))
21:53:47
shinmera
(register-object-for-pass pass (maybe-finalize-inheritance 'trial::lines)))
21:53:52
shinmera
lichat does support it of course.
21:54:06
shinmera
I think weechat just splits it though?
21:54:11
SAL9000
if tag == id: found_id = True
21:54:11
SAL9000
if tag == source: found_source = True
21:54:19
SAL9000
not sure if that was one message or 2 though...
21:54:28
shinmera
use the web client, heh
21:55:14
shinmera
too lazy to enter the password?
21:55:19
SAL9000_
https://chat.tymoon.eu/data//shirakumo/b6ace6b614e62173852ca418714f55720eb1b6820dbf361f6a9b8a08587bc6b8.png
21:55:29
SAL9000
wasn't sure what would happen if I logged in twice
21:55:51
shinmera
I guess weechat does some weird splitting on its own, then.
21:56:29
SAL9000
yeah, using multiline.pl I can send real multiline messages
21:56:39
shinmera
I thought I had that active for pastes...
21:57:37
SAL9000
you can set multiline.pl to ONLY handle pastes, then you don't get the annoying delay
21:58:16
shinmera
Well, stuff to worry about for another day for me
Friday, 11th of June 2021, 3:02:43 UTC