freenode/lisp - IRC Chatlog

7:25:28 #lisp ** TOPIC Common Lisp, the #1=(programmable . #1#) programming language <http://cliki.net/> <http://paste.lisp.org/new> logs:<http://ccl.clozure.com/irc-logs/lisp/>|contact op if muted|SBCL 1.3.0, ASDF 3.1.6, CMUCL 21a, ECL 16.0.0, CFFI 0.16.0, cl-launch 4.1.4, flexi-streams 1.0.15, Hunchentoot 1.2.34, Drakma 2.0.1

7:27:24 #lisp ** TOPIC Common Lisp, the #1=(programmable . #1#) programming language <http://cliki.net/> <http://paste.lisp.org/new> logs:<http://ccl.clozure.com/irc-logs/lisp/>|contact op if muted|SBCL 1.3.0, ASDF 3.1.6, CMUCL 21a, ECL 16.0.0, CFFI 0.16.0, cl-launch 4.1.4, flexi-streams 1.0.15, Hunchentoot 1.2.34, Drakma 2.0.1

7:33:23 pepol-otw ** NICK pepol

7:36:31 Harag1 ** NICK Harag

7:36:59 josteink_ ** NICK josteink

7:40:25 sbryant_ ** NICK sbryant

7:40:25 raeonite ** NICK raeon

7:40:27 profess_ ** NICK profess

7:41:03 otwierac1 ** NICK otwieracz

7:41:04 _death` ** NICK _death

7:41:04 Plas ** NICK PlasmaStar

7:41:04 Byakuren` ** NICK Byakuren

7:41:08 killmaster_ ** NICK killmaster

7:41:33 gz_ ** NICK gz

7:43:24 eazar001 ** NICK rainbowpwny

7:43:56 ozzloy_ ** NICK ozzloy

7:48:15 #lisp ** TOPIC Common Lisp, the #1=(programmable . #1#) programming language <http://cliki.net/> <http://paste.lisp.org/new> logs:<http://ccl.clozure.com/irc-logs/lisp/>|contact op if muted|SBCL 1.3.0, ASDF 3.1.6, CMUCL 21a, ECL 16.0.0, CFFI 0.16.0, cl-launch 4.1.4, flexi-streams 1.0.15, Hunchentoot 1.2.34, Drakma 2.0.1

7:50:23 rainbowpwny ** NICK idkfa

7:51:22 idkfa ** NICK iddqd

7:52:12 MikeSeth_ ** NICK MikeSeth

7:56:04 pepol-otw ** NICK pepol

7:56:19 oGMo_ ** NICK oGMo

7:56:20 seg_ ** NICK seg

7:59:54 _main_ ** NICK __main__

8:00:59 iddqd ** NICK eazar001

8:04:29 swflint_away ** NICK swflint

8:09:18 j0k ** NICK 7YUAAEJZW

8:12:03 tkd_ ** NICK tkd

8:13:59 xristos ** NICK Guest95815

8:25:07 p_l_ ** NICK p_l

8:34:00 mbuf_ ** NICK mbuf

8:34:56 DeadTrickster__ who is in charge of common-lisp.net? how it come I should send email to register? what a nonsense

8:38:27 farhaven_ ** NICK farhaven

8:41:18 alms_clozure_ ** NICK alms_clozure

8:45:32 antoszka_ ** NICK antoszka

9:22:52 loke Hello DeadTrickster_

9:27:28 jackdaniel DeadTrickster__: talk with ehu, or just write to admin@common-lisp.net

9:27:52 DeadTrickster__ loke, hey

9:29:35 DeadTrickster__ jackdaniel, nah just i'll just continue with github

9:30:30 jackdaniel I've just answered you question – continue wherever you want ^_^

9:30:35 jackdaniel s/you/your/

9:30:51 DeadTrickster__ :-)

9:31:14 DeadTrickster__ seriously though, it has login with github option that doesn't work

9:35:05 jackdaniel I'm sure cl.net admin will gladly accept your help with fixing this issue (if you know how to configure gitlab)

9:35:09 gdmalet ** NICK Guest70507

9:37:14 DeadTrickster__ jackdaniel, I wish I had second life just for fixing random CL-related stuff

9:42:32 jackdaniel what's the point of criticizing then?

9:43:27 Shinmera jackdaniel: Getting people aware of problems and issues?

9:43:36 Shinmera *Getting people to be

9:44:47 jackdaniel most direct path is writing to the mailing list or directly to admin. rambling on irc is just rambling

9:45:07 Shinmera In case you didn't notice already, you aren't exactly helping by rambling about his rambling.

9:45:35 jackdaniel I'm talking directly with person who's "admin"

9:45:52 ferada_ ** NICK ferada

9:51:20 abbe_ ** NICK abbe

9:52:11 qubitnerd ** NICK eru

10:11:44 namespace http://jdpressman.com/2015/11/25/how-to-setup-a-common-lisp-web-environment-%28november%202015%29.html

10:11:48 namespace * finally finished this

10:12:17 namespace I'll add a section explaining Woo soon, which is why it's listed at the top.

10:12:22 namespace But otherwise, any critque/comment?

10:19:15 Firedancer namespace: Have these actually been tested on Windows?

10:19:21 namespace Firedancer: No.

10:19:31 namespace Firedancer: I don't have a windows machine handy to do so.

10:20:05 namespace Which is part of why I'm asking here.

10:40:37 wizzomafizzo ** NICK wizzo

11:13:33 PuercoPop pjb: have you considered breaking up your cl reader into smaller API. In the line of the util potential-number-p. stuff like a read-token would be useful for when writing reader macros like the pseudonym's proposed by phoe

11:22:46 phoe_krk pjb: ^

11:24:13 phoe_krk I'd gladly use some better-thought code inside my pseudonyms

11:29:22 splittist beach's Cluffer library would have been useful when I was half-writing triviil ('The Really Incomplete Vi Implementation In Lisp')

12:06:27 nortti_ ** NICK nortti

12:07:03 lnostdal_ ** NICK lnostdal

12:15:59 pjb PuercoPop: yes. time.

12:23:53 DGASAU` ** NICK DGASAU

12:59:32 Shinmera Fun snippet of the day: (defun ​(​) (if (<= 1 ​ 2)​ (* ​(​(1- ​)) 1)))

13:02:06 phoe_krk is it valid if you don't provide a name?

13:02:47 phoe_krk ...

13:03:18 phoe_krk you provide a name

13:03:28 phoe_krk and the function does nothing

13:03:31 Shinmera Nope.

13:03:54 phoe_krk (defun name args &body body)

13:03:56 phoe_krk name: ()

13:04:03 phoe_krk args: (if (<= 1 ​ 2)​ (* ​(​(1- ​)) 1))

13:04:05 phoe_krk body: nil

13:04:18 Shinmera Not quite.

13:04:28 Shinmera Copy paste it into your lisp and see what happens.

13:04:32 phoe_krk Nothing.

13:04:34 phoe_krk It seems.

13:04:41 phoe_krk Just an empty line for me.

13:05:18 jackdaniel phoe_krk: you may try (funcall * 3) for your amusement

13:05:26 jackdaniel after this empty line

13:05:32 phoe_krk dafuq

13:05:39 Shinmera :^)

13:06:18 phoe_krk WARNING: redefining CL-USER::​ in DEFUN

13:06:20 phoe_krk awesome

13:06:23 phoe_krk how do I get rid of this now

13:06:25 phoe_krk xD

13:10:41 Cymew Shinmera: That one was weird...

13:11:15 Shinmera Surprise your friends, confound your enemies!

13:13:25 synchromesh ABCL seems fine with it - (funcall * 3) => 6

13:13:30 synchromesh FWIW

13:13:55 jackdaniel it should, unless it doesn't support unicode characters

13:13:58 Shinmera Well if you have unicode it's perfectly valid.

13:15:12 synchromesh Shinmera: OK, well that explains it then. :)

13:17:46 otwieracz Hey!

13:18:06 otwieracz I've *probably* found bug in ASDF/quicklisp: http://paste.lisp.org/display/160873

13:18:21 otwieracz Could you guys please take a look?

13:18:44 otwieracz (replacing this test-case with IOLIB gave the same results)

13:18:57 DeadTrickster__ anyone is using cl-smtp?

13:19:10 otwieracz XachX: maybe you'll be interested: http://paste.lisp.org/display/160873

13:19:31 H4ns otwieracz: what is the bug that you think you've found?

13:19:45 H4ns DeadTrickster__: i've used it a lot.

13:19:48 otwieracz For me it looks like Quicklisp does not load dependencies for defsystem-dependencies.

13:20:12 otwieracz system X has :defsystem-depends-on :Y

13:20:21 otwieracz And :Y has :depends-on ("alexandria")

13:20:40 DeadTrickster__ H4ns, saw your? -hh fork

13:20:40 otwieracz While quickloading :X, ASDF crashes because there is no alexandria system available.

13:20:54 H4ns DeadTrickster__: mine, yes.

13:20:54 otwieracz (and Quicklisp did not loaded it yet)

13:21:03 otwieracz H4ns: There are steps to reproduce in this paste.

13:21:22 DeadTrickster__ I just push "flexible attachments" https://github.com/deadtrickster/cl-smtp/tree/flexible-attachments branch - now one can subclass attachments and I want feedback

13:21:43 DeadTrickster__ I also implemented stream-attachment and base64-attachment

13:22:09 DeadTrickster__ I plan to use latter with files-stored-in-db

13:22:26 H4ns otwieracz: https://bugs.launchpad.net/asdf/+filebug

13:22:43 otwieracz H4ns: I am not sure if this is not quicklisp bug.

13:23:13 otwieracz Quicklisp should load defsystem dependecies dependencies before letting ASDF load dependant systems.

13:23:58 DeadTrickster__ H4ns, also, should I look at your branch closely if cl-smtp just works?

13:25:18 H4ns DeadTrickster__: no. i've made several incompatible changes that did not quite follow the follow the philosophy of cl-smtp, which is why the fork exists. i've wanted to be able to format mail bodies myself, whereas cl-smtp tries to do that for the user.

13:25:41 H4ns DeadTrickster__: your changes seem like a good addition to the main line, not to my fork.

13:26:10 DeadTrickster__ thanks, want you mean by 'format mail bodies'?

13:26:52 H4ns DeadTrickster__: format mime message in application or with other libraries, leave only the sending to cl-smtp

13:27:27 DeadTrickster__ like separating serialization and transport protocol, right?

13:27:45 DeadTrickster__ how this can be useful? could you give an example?

13:28:08 H4ns DeadTrickster__: right. at the time, i had very specific requirements how i wanted the mime message to look like, i.e. with inline images, alternate bodies etc.

13:28:34 JuanDaugherty * used BEE

13:29:04 H4ns DeadTrickster__: this was not something that could easily be generalized.

13:29:04 DeadTrickster__ sounds like mime streams )

13:29:23 phoe_krk What's the easiest way of removing the last character off a string?

13:29:38 DeadTrickster__ H4ns, thanks, I'm simple guy, cl-smtp is enough I hope )

13:29:58 H4ns DeadTrickster__: probably so

13:30:01 JuanDaugherty phoe_krk, don't remove it just stop thinking about it

13:30:07 phoe_krk JuanDaugherty: ldsjghsklghdflg

13:30:36 phoe_krk What's the easiest way of returning a string without its last character?

13:30:38 DeadTrickster__ H4ns, I just hope it won't interfere with content like some of cl-sendmail variants

13:34:43 phoe_krk answer: (subseq string 0 (- (length string) 2))

13:35:00 Shinmera Beware of empty strings.

13:35:29 phoe_krk Yes, I take that into account before.

13:35:50 DeadTrickster__ ** NICK DeadTrickster

13:36:04 phoe_krk actually

13:36:12 phoe_krk (subseq string 0 (- (length string) 1))

13:36:39 Shinmera There's 1- for that.

13:36:50 phoe_krk Oh, correct.

13:45:48 sz0 ** NICK sz0`

13:59:35 sz0` ** NICK sz0

14:18:14 pjb phoe_krk: notice that subseq copies. You might want nsubseq.

14:18:35 phoe_krk Does nsubseq destroy the string it works on?

14:18:37 pjb com.informatimago.common-lisp.cesarum.utility:nsubseq

14:18:41 pjb Nope.

14:18:54 Shinmera On the other hand indexing with a displaced array can be very slow. More so than the overhead of copying the string in many cases.

14:18:55 pjb But it is Non-consing

14:19:06 pjb (well minimally consing).

14:19:49 pjb Shinmera: not slower than indexing an array in C (which is always done thru the indirection of a pointer).

14:19:56 phoe_krk well, the strings I work on here aren't going to be longer than a few characters.

14:20:12 Shinmera pjb: Type inference suffers greatly from displaced arrays, which can make things very slow.

14:20:12 pjb Good.

14:20:22 pjb Shinmera: only with bad implementations like sbcl.

14:20:23 phoe_krk I can pay the price for using subseq.

14:21:30 Shinmera pjb: You might want to talk to someone about your consistent hatred for sbcl.

14:21:38 pjb phoe_krk: yes, for small strings. Also, displaced arrays keep a reference on the original array so if you start from big arrays and only use small part, that would prevent garbage collection of the unused parts.

14:21:59 pjb Shinmera: when people will talk to someone about their consistent hatred for clisp.

14:22:40 Shinmera Taking "revenge" on implementations is hardly a sane train of thought.

14:22:52 pjb Shinmera: the problem is of course not with sbcl, but with programmers who use sbcl and who retain the same reflex as C programmers: only thinking about speed.

14:33:25 pyon ** NICK quater-nyan

14:46:59 phoe_krk I have this macro, http://paste.lisp.org/display/160883, and later I have (defun a () (alist-gen :q :w)) (defun b () (alist-gen (:z (a))). When compiling that, it throws The function LISPFURC::A is undefined.

14:47:15 phoe_krk Well, that's problematic, I define it just above.

14:47:38 phoe_krk Or maybe the problem is there within that eval call.

14:47:38 DeadTrickster displaced arrays are really good idea, however I'm always worried about performance and gc and basically never use them

14:48:02 DeadTrickster and yes I'm primarily sbcl guy

14:49:09 DeadTrickster Shinmera, how type inference suffers?

14:49:23 DeadTrickster element-type is known, bounds are known too

14:49:28 DeadTrickster what i'm missing?

14:50:11 Shinmera DeadTrickster: The type system doesn't support distinguishing them, so you have to test all the time.

14:50:33 DeadTrickster distinguishing between displaced arrays?

14:51:05 Posterdati hi

14:51:16 Shinmera Well for one there's no displaced-array type.

14:52:10 DeadTrickster is it really needed? it can inherit many things from base array i.e. vector/basic-vector

14:52:58 DeadTrickster or speaking of sbcl displaced array for unsigned-byte 8 simple-array will be just a matter of adjusting pointers

14:53:01 Shinmera I'm not the best to ask about this because I don't know the internals and don't remember the discussion in full. I'm sure someone on #sbcl like stassats could explain better though.

14:53:22 DeadTrickster yeah stassats )

14:53:30 Posterdati how can I print a binary number with format with a variable number of digit depending on a variable?

14:53:30 DeadTrickster he is always there

14:54:25 Cymew pjb: You use clisp?

14:54:29 phoe_krk Can I anyhow modify the order of macroexpansion and compilation?

14:54:55 phoe_krk Other than on a per-file basis?

14:56:29 sz0 ** NICK sz0`

15:00:15 Cymew pjb: The reason I'm curious is how its CLOS implementation is. I've gotten the impression there was something wrong with it, but I don't know any more than that.

15:03:47 phoe_krk Uh. How do I tell Lisp to compile one expression before macroexpanding another?

15:03:57 White_Flame Posterdati: (format nil (format "...insert numeric literal..." digit-length) number) comes to mind

15:04:09 White_Flame ...format nil on the inner, too

15:04:54 White_Flame phoe_krk: put it in a file that's loaded earlier via asdf?

15:05:27 phoe_krk White_Flame: that's a solution, but I'd suddenly need to pop in four different files just to make my simple thing work =_=

15:06:12 White_Flame have you tried eval-when?

15:06:42 phoe_krk trying now

15:06:51 White_Flame and the dirtiest hack is #.(make-it-happen-at-read-time)

15:07:17 phoe_krk Jesus, that's dirty

15:08:43 Posterdati White_Flame: tx

15:10:34 rszeno Posterati, (format t "~v,'0b" digit-length number)

15:10:54 rszeno Posterdati, ^^

15:11:01 phoe_krk (eval-when (:compile-toplevel) ...) worked. Thanks!

15:11:10 phoe_krk Just what I was looking for, too.

15:11:28 Posterdati rszeno: ah ~v ! Thanks

15:12:07 Cymew Feel free to call me stupid, but what's wrong here: (defmacro get-list-item (lst key `,(aref lst key)) and (get-list-item (make-array 10) 1) claims its input is not an array?

15:12:33 Cymew I'm kind of fried so this is probably obvious...

15:12:36 White_Flame (lst key) ?

15:13:16 White_Flame `(aref ,lst ,key)

15:13:18 Cymew oops, that's a paste error. I should have used pastebin...

15:13:29 Cymew Hmm. Can you have multiple commas?

15:13:42 White_Flame sure, it nests

15:14:49 White_Flame what's happening here is that at compile time, the macro body is _evaluating_ the equivalent of (aref '(make-array 10) 1), and '(make-array 10) is not an array

15:14:57 White_Flame (in the bugged case)

15:15:06 Cymew Hmm. That made the error go away, but still did not give me what I expected. I think I should probably go home and sleep.

15:15:11 Cymew Thanks White_Flame

15:15:17 White_Flame in my case, the macro body is constructing & returning the _list_ (aref (make-array 10) 1)

15:15:59 Cymew I've been on the verge of using EVAL a couple of times now, so I'm bad off

15:16:20 White_Flame what expectation was violated?

15:17:30 Cymew I tried to create multiple SDL Surfaces and have an array as a lookup list for them, but I didn't get my correct Surface out.

15:18:10 Cymew Ned to re-think this

15:18:41 White_Flame ok, so that goes beyond the macroexpansion

15:33:23 phoe_krk Within iterate, how do I iterate on two variables by once?

15:33:28 phoe_krk Like, I want to iterate over a plist.

16:13:46 hellekin hello, is anyone involved with gcl? I need a gcl maintainer to fix a link on their website.

16:22:16 Cymew phoe_krk: Something like this? (LOOP :for p from 1 to 10 :for q from 1 to 10 :do (FORMAT t "~A ~A - " p q))

16:22:27 phoe_krk Cymew: using iterate

16:22:37 phoe_krk and over a plist

16:22:39 Cymew Not ITERATE, obviously, but LOOP can do it so...

16:22:58 phoe_krk if loop can do it, so must iterate

16:23:24 Cymew Possible. I don't use ITERATE myself, and I think you'll find more LOOP knowledge around.

16:23:47 Cymew But, then I at least got what you meant, and I'm not as fried as I thought. Always something.

16:25:13 Cymew Iterate doc strings really sucks.

16:28:22 cyraxjoe_ ** NICK cyraxjoe

16:35:31 phoe_krk holy cow, imperative programming becomes so fun with the use of closures, too

16:35:37 stepnem_ ** NICK stepnem

16:36:00 cataska_ ** NICK cataska

16:36:11 Shinmera- ** NICK Shinmera

16:47:30 vydd` ** NICK vydd

17:32:13 stanislav ** NICK quasus

17:42:19 pjb Posterdati: (format nil "~V,'0B" 8 42) #| --> "00101010" |#

17:47:11 khisanth_ ** NICK Khisanth

17:50:07 Posterdati pjb: yes, thanks

18:02:13 nopf_ ** NICK nopf

18:32:36 KyoriAsh__ ** NICK KyoriAsh

18:39:07 axion how can i make a random-state object such that each call to random produces a series of predictable values?

18:39:20 Shinmera You cannot.

18:39:31 Bicyclid1ne er, predictable in what sense?

18:39:36 Bicyclid1ne ** NICK Bicyclidine

18:39:54 axion i'm procedurally generating game worlds, and i'd like to save the seed used so i can generate the same world again

18:41:29 shka hmmm

18:41:31 shka good question

18:41:39 Shinmera The random-state is a printable object.

18:41:40 Bicyclidine yeah, sure. just save *random-state* at some point. if you start generating randoms with the same random state you should get the same sequence.

18:41:48 shka perhaps dynamic variable?

18:41:51 shka riiight

18:41:55 Shinmera But you cannot seed it by the user, as the internals are implementation dependant.

18:42:10 axion right

18:42:26 Shinmera I've started writing a library of portable RNGs for that reason, but never got it done. Not enough time in the day!

18:43:04 Bicyclidine you can give it some arbitrary seed with (make-random-state t). and some implementations have extensions, so you can do e.g. sb-ext:seed-random-state.

18:43:14 Bicyclidine i think there's also a portable mersenne twister somewhere.

18:43:18 axion to be clear, the same implementation will be used by all users

18:43:57 Bicyclidine minion: MT19937?

18:43:57 phoe_krk Is there a way for the printer to acknowledge the available line width?

18:43:57 minion MT19937: No definition was found in the first 5 lines of http://www.cliki.net/MT19937

18:44:10 phoe_krk Not the reader, the REPL

18:44:24 Bicyclidine clhs *print-right-margin*

18:44:24 specbot http://www.lispworks.com/reference/HyperSpec/Body/v_pr_rig.htm

18:44:25 Bicyclidine that?

18:44:28 phoe_krk I'm using slime and I'd like the output to take the whole width of thanks

18:45:41 KyoriAsh__ ** NICK KyoriAsh

18:46:01 Bicyclidine huh. long array gets cut off, you're right. i did not realize.

18:46:06 sweater ** NICK Guest61627

18:46:25 phoe_krk precisely.

18:46:30 phoe_krk Bicyclidine: long lists, too.

18:46:46 phoe_krk But I'm using a listy data structure, so I see that sort of thing.

18:47:47 Bicyclidine i think i do not know why that is happening...

18:48:06 Bicyclidine oh, it's part of the pretty printer.

18:51:56 gilez_ ** NICK Guest57119

19:01:55 seg_ ** NICK seg

19:01:55 themattchan ** NICK mach