Search
Sunday, 20th of May 2018, 5:56:35 UTC
6:35:12
vtomole
clhs integer-length
6:35:13
specbot
http://www.lispworks.com/reference/HyperSpec/Body/f_intege.htm
6:41:33
vtomole
clhs array-dimension
6:41:34
specbot
http://www.lispworks.com/reference/HyperSpec/Body/f_ar_dim.htm
12:16:50
makomo
how do i leave a note to someone using Colleen again?
12:17:22
pjb
Oh, colleen, another bot I didn't know.
12:17:44
makomo
i wasn't aware of minion :-)
12:18:37
pjb
Colleen is not particularly helpful…
12:18:37
Colleen
Unknown command. Possible matches: 8, set, say, mop, get, hello, grant, time, tell, roll,
12:19:25
Colleen
pjb: Unknown command. Possible matches: 8, tell, set, get, time, help, deny, say, mop, test url,
12:19:59
makomo
oh, "test!" must be a command, can't be an arbitrary string
12:20:20
makomo
ah, here it is, ::notify
12:20:51
pjb
You have to prefix all commends with ::tell? /msg Colleen ::tell help
12:21:53
makomo
no, the syntax is "::<command-name> <args>"
12:22:33
makomo
::tell runs another Colleen command and shows/tells it to someone else
12:22:33
Colleen
runs: Unknown command. Possible matches: 8, weather, set, say, mop, get, notify, grant, block, award,
12:22:51
makomo
and ::notify just sends someone a message the first time they speak again
13:15:06
ebzzry
Is usocket *the* cross platform TCP/IP solution? I want to port a LispWorks library to use it.
13:18:47
White_Flame
it's quite high in use: http://blog.quicklisp.org/2018/03/download-stats-for-february-2018.html
13:29:59
ebzzry
the next question for myself, is what are the equivalent function from lispworks.
13:43:52
White_Flame
ebzzry: looking at usocket's lispworks backend, it looks like they're all in the comm: package https://github.com/usocket/usocket/blob/master/backend/lispworks.lisp
13:48:01
oleo
does anyone know the error relating to the last link i provided
13:48:12
oleo
the former link is one where i dump an image from sbcl
13:48:48
oleo
it all goes ok, the image gets dumped but on start i get library error
13:49:31
oleo
the whole asdf mockery is because i have two copies of mcclim
13:49:49
oleo
mcclim-master and mcclim-freetype2, and the error is related to freetype2 library
13:50:15
oleo
i basically am dumping the image while ignoring the master
14:18:08
drmeister
Should (find-class '(setf foo) nil) return NIL or signal an error?
14:18:30
Bike
it's undefined, but an error would be nice
14:18:48
drmeister
An error breaks manifest.
14:19:13
drmeister
Clasp signals an error and that breaks manifest
14:19:27
Bike
i'd say that's unfortunate on manifest's part
14:20:39
drmeister
https://github.com/quicklisp/quicklisp-client/blob/master/dists/quicklisp/software/manifest-20120208-git/manifest.lisp#L274
14:21:03
drmeister
It doesn't test that **symbol** is a symbol.
14:25:16
specbot
http://www.lispworks.com/reference/HyperSpec/Body/f_find_c.htm
14:25:39
phoe
should always return an error
14:26:02
phoe
the variable "symbol" is meant to be a symbol and therefore calling FIND-CLASS with a non-symbol invokes undefined behavior
14:26:34
phoe
I mean, the spec says that it may return an unicorn because the behavior is undefined in such a case
14:27:37
phoe
drmeister: also, your link gives me 404
14:28:35
drmeister
I've never figured out how to get those links to work with quicklisp packages.
14:32:32
phoe
drmeister: https://github.com/gigamonkey/manifest/blob/master/manifest.lisp#L274
14:34:11
phoe
drmeister: how come that you end up calling FIND-CLASS with a non-symbol though?
14:34:40
Bike
because manifest just iterates over all symbols and all (setf symbol)s, probably./
14:35:28
phoe
I see. Then it's manifest's fault for invoking UB in the first place if it calls FIND-CLASS on non-symbols.
14:37:11
drmeister
That's what appears to be happening - calling find-class on non-symbols
14:38:36
drmeister
https://github.com/clasp-developers/manifest/blob/master/manifest.lisp#L274
14:39:35
phoe
(and (symbolp symbol) ...) should do it if I read this macro correctly
14:40:05
phoe
the :IS form seems like a filtering predicate that checks whether a thing falls into a category
14:51:57
drmeister
Yes - and it needs to be added to the :is test for :condition as well
14:52:16
phoe
Yep, since condition types are found the same way.
14:54:44
drmeister
Now manifest works with clasp
15:03:05
phoe
drmeister: remember to make a PR to the mainstream repo!
15:42:41
oleo
http://dpaste.com/22F1R3C http://dpaste.com/1QRCND6, i did a backtrace for the last one but have still no clue what's wrong
15:45:16
pjb
oleo: typing v on the line 0: in the backtrace should jump to the source line, and you should see where the library value comes from. It's probably the variable that's bad, according to the error message.
15:46:09
pjb
oleo: IIRC, typing RET on that line should show the frame with the bindings too.
15:47:37
beach
oleo: Do you have the right values for your OPTIMIZE qualities?
15:54:15
oleo
beach: no idea beach, do you think it's related to that ?
15:56:15
beach
I just think you need the right values to debug the problem properly.
16:10:47
oleo
and #.(declaim (optimize (safety 3) (debug 3) (space 0) (speed 0) (compilation-speed 0) (inhibit-warnings 0))) is not good enough ?
16:32:58
beach
Another FFI-based system, I guess?
16:39:59
on_ion
well freetype2 is a C lib..
16:41:42
oleo
and 5am gives some results like
16:57:20
oleo
but if i were to speculate, it either is some loop gone wrong, types not fitting or some conversion stuff in between.....
Sunday, 20th of May 2018, 17:56:35 UTC