freenode/lisp - IRC Chatlog

6:32:37 beach Shinmera: The URL http://log.irc.tymoon.eu/freenode/lisp works, but http://log.irc.tymoon.eu/freenode/clasp does not. Any reason for that?

6:33:00 beach With the latter, one has to add arguments in order for it to work.

6:33:29 Shinmera Both work fine for me

6:34:12 beach For the latter one, I get "Secure Connection Failed"

6:34:28 Shinmera did you put in another https by accident on your end?

6:34:40 Shinmera because the links you pasted me just now both work.

6:34:50 beach Wait...

6:35:15 beach I said the wrong thing...

6:35:31 beach Shinmera: The URL log.irc.tymoon.eu/freenode/lisp works, but log.irc.tymoon.eu/freenode/clasp does not. Any reason for that?

6:35:58 Shinmera your browser caching one as https and the other not.

6:36:07 Shinmera Apparently anyway, if that's the error you're getting.

6:36:09 beach But actually, it is still true for the first two URLs.

6:36:39 Shinmera Again, I cannot reproduce your issue, but if you're getting an error about secure connections, then your browser is trying to establish HTTPS, which will not work.

6:36:48 beach Right.

6:36:54 beach I don't know why it is doing that.

6:37:19 Shinmera The only thing I can think of is an extension like https everywhere that tries to do that, or your browser caching it and trying to be smart.

6:37:35 beach Even though I give an explicit http: it still tries https.

6:37:44 Shinmera Either way, I should probably move the logs down to a single subdomain level so https will actually work.

6:37:48 Shinmera Huh.

6:38:39 beach Even though I give http://log.irc.tymoon.eu/freenode/clasp it displays the URL with https and fails.

6:39:04 beach Now, if I put in explicit arguments, it works.

6:39:11 beach like &from &to etc.

6:41:32 Shinmera Clearing your cache might be an option.

6:41:46 beach OK, I'll try that. Thanks.

6:45:58 Shinmera This should work now as well: https://irclog.tymoon.eu/

6:46:08 Shinmera (the https is intentional this time)

6:46:54 theos hey

6:48:15 theos what would be the best approach to make a massively modular CL system? i want children (mostly my kids) to develop it as they grow up.

6:49:00 beach minion: Please tell theos about SICL.

6:49:00 minion theos: SICL: SICL is a (perhaps futile) attempt to re-implement Common Lisp from scratch, hopefully using improved programming and bootstrapping techniques. See https://github.com/robert-strandh/SICL

6:50:21 theos beach i want to use CL as it is.

6:50:47 beach theos: I am confused.

6:51:06 beach Perhaps you mean something different than I do when you say "a ... CL system".

6:51:41 theos hmm. so CL cant be used to make massively modular systems/software?

6:51:43 beach Are you talking about an application written in CL?

6:51:49 theos yes^^

6:52:02 beach We mean different things when we say "a ... CL system" then.

6:52:11 beach OK, now I understand.

6:52:22 theos what do you mean when you say that?

6:52:31 beach an implementation of CL.

6:52:36 beach That's what I thought you meant.

6:53:02 beach Like "a Unix system" is not an application that runs on Unix, but an implementation of Unix. To me, that is.

6:53:08 theos oh. i should have used "software"

6:53:11 beach But now I understand.

6:53:23 beach My advice, use CLOS a lot.

6:53:30 White_Flame "system" as used in ASDF

6:53:38 beach True.

6:53:39 theos should i read the amop book?

6:53:54 beach theos: No need if you are just going to use CLOS.

6:54:12 beach theos: But you need to understand it fully as a user.

6:54:28 beach Unfortunately there are few books about it, and they are a bit old.

6:55:15 beach theos: Define protocols for each part of your software.

6:55:24 theos beach can you suggest some(or all of them if thats ok with you)? i might have a few

6:55:58 beach theos: Here, a "protocol" is a set of generic functions and protocol classes.

6:57:06 beach The only book I know of really is Sonja Keene's book.

6:57:36 beach But I never quite understood CLOS from that book. I understood it from reading the CLIM II specification.

6:57:39 theos ah! i have that one. i think thats the best one written on CLOS

6:58:00 beach It's the best one, but it's not good. :)

6:58:23 theos not good how?

6:58:25 beach The examples are old and confusing.

6:58:52 beach The first example uses locks, which is not a standard Common Lisp feature.

6:59:24 theos hmm so its an old book even for CL standards

6:59:57 beach In my opinion it is old, yes. And very incomplete.

7:01:04 theos CLOS is introduced in a lot of CL books as a chapter or 2. i think that would be enough.

7:01:15 beach These days, I tend to break up my programs into "modules", where each module has an ASDF system definition, a package, one or more protocols, and one or more files.

7:03:18 beach If you do that, you make sure that each module is independent of the program to which it belongs, and it could potentially be used in other programs as well.

7:04:16 beach Do you have some particular application in mind?

7:05:15 theos yes thats what i am planning to do. so each module can be removed and added without affecting the application.

7:05:30 beach Right. That's a good start.

7:08:54 beach For each protocol, you establish, document, and enforce preconditions and invariants.

7:10:34 theos i am thinking of making an educational software for my kids and other kids that i teach. it should be engaging and interesting enough that they will learn everything through it and also learn CL by hacking it. modifying old modules, adding new ones as they learn more etc. the "system" will be a combination of a lot of systems which can be used independently and also together.

7:12:23 theos it will teach kids and also learn how to teach as time progresses

7:13:29 beach For each module, define specific condition types to be signaled, as well as restarts whenever they make sense, and document those as part of the protocol.

7:14:12 beach Interesting project.

7:14:16 beach Hard stuff though.

7:14:34 theos yes there will be an inference engine and a knowledge base in one independent part of the system.

7:16:47 beach If you are not very experienced with CLOS, I suggest you submit your modules for review here, at least in the beginning.

7:18:20 beach Have you already started writing some code?

7:19:05 theos not yet. i am looking for the best approach

7:20:06 beach I am willing to bet that you need a very different approach for different subjects. I don't see much in common between (say) mathematics and geography.

7:21:44 beach I am assuming you want the user to answer questions asked by the software and that you want to be able to check the answers.

7:21:56 aap_ ** NICK aap

7:25:54 theos yes question/answer expert system is one part. the main part will be the teaching system. like a hypermedia wiki where users can learn about anything they want. the course/syllabus will be constructed separately for every user after the system knows how much or how little the user knows and wants to know.

7:27:06 theos i want to make the ultimate learning environment.

7:31:56 guthur there was some expert system teaching platform, written in Lisp iirc

7:32:09 guthur can not remember the name right now

7:38:47 theos oh nice! i will search for it

7:43:00 fe[nl]ix ** 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| ASDF 3.1.5, cl-launch 4.1.4, SBCL 1.2.13, flexi-streams 1.0.15, Hunchentoot 1.2.34, Drakma 2.0.1

8:17:58 jackdaniel * -- the bughunt :D

8:52:07 jackdaniel oh my, never tought I'll enjoy daft punk as theme for programming ^_^

8:56:27 clique Music has kept becoming lower-class because the demographics of people who can afford to patronize music has kept extending to lower classes

8:57:10 schjetne I'm not sure I would call subscribing to Spotify "patronage"

8:57:16 Shinmera Fortunately this is #lisp and not #music

8:58:30 clique xd

8:58:32 jackdaniel and to make it on-topic - all programming theme is good for me is in CL ;-)

8:59:24 clique one more time xd

10:17:03 jackdaniel what may cause, that #\A is read by reader as #\a ?

10:17:29 jackdaniel (any character downcased that is)

10:17:46 jackdaniel something with readtable may cause that?

10:18:24 oleo_ *print-case* ?

10:19:01 oleo_ that's for printing and string-downcase can do that too

10:19:34 jackdaniel yes, but I'm asking about reader

10:20:04 jackdaniel (read) ;; #\A ;; --> #\a

10:20:26 oleo_ (readtable-case *readtable*)

10:20:32 oleo_ that one too

10:20:48 oleo_ it's :upcase in my repl

10:20:55 oleo_ and normally it is :upcase afaik

10:21:07 jackdaniel yes, but it doesn't affect character literals, like #\a #\A

10:21:15 jackdaniel it affects symbols

10:24:42 PuercoPop (read #\A)

10:24:48 PuercoPop wrong window

10:57:45 beach` ** NICK beach

11:09:08 theos hey

11:11:23 jackdaniel o/

11:13:37 theos * listens to "its my life" while writing CL

11:55:50 magistr give me a Common Lisp HyperSpec in pdf format

11:58:23 guthur magistr not sure it would be a "hyper"Spec then, there is the ANSI standard PDF

11:58:53 guthur * has not actually looked at the PDF version of the ANSI spec

11:59:01 k-stz magistr: if you need it offline there are better solutions

11:59:04 guthur there is a link from http://www.lispworks.com/documentation/common-lisp.html

11:59:58 magistr "The document you tried to access is not available or may have been replaced by a newer version. "

12:00:31 jackdaniel guthur: ansi pdf quality is terrible

12:00:38 jackdaniel at least that's what i've heard

12:01:23 trig-ger_ ** NICK trig-ger

12:04:08 magistr " The requested page "/Web/Groups/AI/html/cltl/cltl2.html" could not be found."

12:05:10 remi`bd magistr: http://www-prod-gif.supelec.fr/docs/cltl/cltl2.html

12:06:20 knobo Would it be possible to make hunhentoot fall back to latin-1 if it got an utf-8 charachter error?

12:06:54 knobo by retrying the whole operation with *hunchentoot-default-external-format* set to +latin-1+

12:10:21 knobo So in stead of crashing the server if someone puts an ø the url, it could give a propper 404

12:14:44 knobo by automatically running recompute-request-parameters on a decode-error

12:29:47 jackdaniel o, mesa hit 4.0

12:29:55 jackdaniel opengl 4.0 that is°

12:30:05 |3b| nice

12:32:03 k-stz thanks valve I guess

12:32:13 |3b| or just time

12:33:04 jackdaniel they strived for it for quite a long time

12:37:09 k-stz jackdaniel: source?

12:37:41 |3b| seems pretty obvious that they would be working on it

12:39:20 jackdaniel k-stz: http://cgit.freedesktop.org/mesa/mesa/

12:41:23 jackdaniel quick ducking returns this http://www.phoronix.com/scan.php?page=article&item=mesa-opengl-40&num=1 , if you want some press release

12:42:25 theos i dont think they can make it better than the proprietary drivers

12:43:11 EdanCoriiin ** NICK EdanCorin

12:46:49 jackdaniel there are many factors which might be associated with "better" - speed is one, another one is interoperability or extensibility, backward compatibility etc etc

12:50:34 k-stz with the 4.0 support for us programmers this just means that we can can write 4.0 programs and expect most linux users to be able to run them right?

12:50:49 jackdaniel yes

12:51:50 edgar-rft magistr: There is a PDF version the dpANS3 pre-print version of the ANSI specification under http://linux4u.jinr.ru/pub/misc/symbolic/lisp/dpANS3/book.pdf

13:21:58 KingNato_ ** NICK KingNato

13:30:08 |3b| k-stz: people will still need 4.x hardware unless you want things to be horribly slow

13:30:38 |3b| (assuming mesa gives you a software fallback instead of just rejecting requests for a 4.x context)

13:36:33 p_l |3b|: well, depends whether the new functions require new hardware (this is not always the case)

13:36:55 |3b| p_l: pretty much the entire point of 4.x is new hardware

13:37:40 |3b| 3.x and 4.x are both actively updated, 3.x is dx9 hardware, 4.x is dx10 (or something like that, i don't remember dx versions)

13:37:58 |3b| though 4.x is pretty old "new hardware" at this point

13:39:25 p_l |3b|: Just more of the "new version mandating what was previously extensions" stuff

13:39:47 |3b| ?

13:40:54 p_l |3b|: Remember the AZDO stuff, it being mostly "oh well, it works if you have 4.3 or so"? Got all of the features on OpenGL 3.3 (except bindless, iirc, which is a hw change) - it was just part of non-mandatory extensions

13:41:33 p_l anyway, Mesa with 4.0 probably means intel got around to updating its hardware to 4.0 ;)

13:55:42 jackdaniel ok, I've found a culpirt, ECL mishandles readtable-case set to :invert *and* readtable isn't reset when loading set of libraries (not sure if latter is problem with ecl, or asdf | quicklisp)

13:57:50 Skrylar i seem to recall ecl is generally unpleasant to work with, although its been a while since i used it and it might be better now

13:58:40 jackdaniel Skrylar: my impression is the oposite, codebase is really clean. Unless you mean end-user experience - slime backend needs a bit of love, that's for sure

13:59:52 Skrylar jackdaniel: my first experiences with it were copious segfaults, found out that for some reason line endings for some random c files had to be set to specific encodings to fix, and it tended to generate code clumsily

14:00:09 Skrylar like, it doesn't actually obey inlines

14:00:55 jackdaniel uhm, I'm not in position to deny what you experienced

14:01:35 jackdaniel anyway I try to squash bugs with time, next time if you try ecl and encounter such a bug, please report it

14:01:50 jackdaniel :)

14:02:19 echo-are` ** NICK echo-area

14:06:13 dim I tried ecl only once and couldn't play with pgloader on ecl, because of lacking of threads in some sorts IIRC, and lparallel or something else not being supported

14:06:31 dim very fuzy memory, only thing I remember is I couldn't run pgloader on ecl

14:07:47 jackdaniel * types in ecl (ql:quickload :pgloader)

14:08:34 jackdaniel cl-postgres-error throws an error about invalid :double-float specializer, hmm. I'll take look into that, if fixing readtables bug succeeds

14:09:07 dim ok, thx for looking into it!

14:09:53 jackdaniel np, I find it really enlightening when lurk through libraries to find bug in them or in ecl - and fix it :)

14:09:55 jackdaniel great learning experience

14:10:30 dim if you play with pgloader you might want to run the unit tests and discover problems, you need some PostgreSQL installed for that

14:10:52 dim pgloader produces and compile lisp code at runtime

14:11:01 dim so chances of new bugs found by running it isn't 0

14:11:13 dim (pgloader:run-commands "/Users/dim/dev/pgloader/test/archive.load")

14:11:22 dim that's how to run a test from the lisp side of the world

14:12:52 jackdaniel hm, when I finish with current thing (:invert), direct goal would be successful pgloader loading into image

14:13:37 jackdaniel I'm not sure if I'll find time to test it throughfully in near future

14:17:20 Skrylar dim: ecl does support threads from what i remember but its very borky

14:17:54 jackdaniel it does, not sure what do you mean by borky tough

14:18:00 Skrylar from what i recall of the SLIME docs, ecl's compiler isn't actually thread-safe which is why slime has to use some of the worse ways of communicating

14:18:36 Skrylar or at least thats what the ecl adapter that shipped with slime said

14:20:41 jackdaniel dim: lparallel and lfarm work fine btw

14:34:42 dim jackdaniel: awesome

14:35:03 dim jackdaniel: as soon as pgloader loads I can handle the testing

14:35:49 dim I will be interested into two things: a. building an image for shipping, with all the .so mess from cffi and b. performances compared to sbcl

14:56:39 m_zr0 man. does anybody have any idea why slime loves to give the error: Symbol's func def is void: slime-face-inheritance-possible-p? I'm trying to learn to be able to set up slime without hassle, and this one plagues me. ELPA install: slime, slime-repl

14:57:15 m_zr0 it's been three weeks (and now we all have a song stuck in our head, at least, the intro)

14:57:51 m_zr0 this happens in windows (solved, don't install slime-repl), linux (WIP), and OSX (wip)

14:58:28 jackdaniel m_zr0: best choice would be using slime took from quicklisp, it's what most people do, so it will be more likely, that they will be able to reproduce and/or you won't have problem

14:59:34 m_zr0 oh nice, looking at it now. I guess i've been doing it the hard way :D

15:00:20 m_zr0 almost makes me sad. you'd think an elpa install would do all of that

15:02:30 jackdaniel good it's only "almost", not "completly" ;-)

15:03:17 m_zr0 ok, this is going to be the biggest beginner question of all time. "load quicklisp into your comm. lisp impl., then eval". does that mean literally download -only- slime-helper.lisp, then load, then run function?

15:03:35 m_zr0 hahah nah the most rewarding things in life are the most challenging to obtain, just an opinion

15:04:14 jackdaniel hm, do something like : sbcl --load quicklisp.lisp

15:04:21 jackdaniel then you'll have a hint message in your repl

15:04:53 jackdaniel how to install quicklisp, and then there is afair a hint, that you want to quickload slime-helper

15:04:53 m_zr0 sorry for the silly questions, really want to (understand) how to get this going. it works on a few systems, but that's like kicking a TV until it works. you should probably just know how you fixed it.

15:05:59 pjb alternatively sbcl RET (load #P"~/quicklisp.lisp") RET

15:06:18 m_zr0 that makes sense, i HAVE to ask. dude this book "land of lisp" insists at the beginning to use clisp, though EVERYONE is using sbcl. will it mess up my (hi pjb!) flow to use sbcl instead of clisp?

15:06:25 pjb there are several ways to do it, that's why we don't tell, we just say "load it".

15:06:43 pjb Nope, I use mostly ccl for now (since I'm working on MacOSX).

15:06:45 |3b| m_zr0: you will need to change some code to use sbcl, but google should be able to find the fixes

15:06:46 jackdaniel m_zr0: I've heard a lot of good things about ccl also

15:06:48 m_zr0 i find way more instructions on sbcl setup, but want to follow the book so i don't run into some nonsense later

15:06:57 pjb usually, we use several CL implementations, since they're good at different things.

15:07:14 |3b| it mostly uses portable CL, but a few things like networking and running programs as child processes differ

15:07:15 pjb sbcl is good at generating slowly fast code, so it may be useful to produce the final executable.

15:07:28 pjb But other implementations may be preferable while developping.

15:08:02 jackdaniel pjb: why do you use macosx? tought you are foss diehard, who despies apple

15:08:05 m_zr0 oooh. so that's why lispers have so many diff lisps on their machines

15:08:11 pjb If you want to run your CL code inside an application, then ecl or abcl might be preferable, depending on whether that application is written in C or java.

15:08:16 pjb Yep.

15:08:30 m_zr0 this is going to be fun. got all tingly inside :D

15:08:33 pjb I use clisp for my scripts since it's a small implementation with an interpreter.

15:08:39 pjb (so no startup time).

15:10:46 m_zr0 woooo brew install sbcl and let's make magic

15:11:02 m_zr0 dude why doesn't osx have wget already geez

15:11:27 pjb m_zr0: it has it.

15:11:28 m_zr0 lol brew install wget....

15:11:39 pjb /opt/local/bin/wget

15:11:41 m_zr0 it does? out of the box ?

15:11:42 pjb port install wget

15:11:48 m_zr0 port.

15:11:55 m_zr0 cmd not found.

15:12:07 m_zr0 since mavericks and newer?

15:12:27 m_zr0 i'm suspect of this install. it seems to be missing all kinds of common programs.

15:12:35 pjb http://macports.org

15:12:45 m_zr0 like, all the terminal, none of the features

15:12:58 pjb And also: http://www.emacsformacosx.com

15:13:38 pjb and http://www.secretgeometry.com/apps/cathode/ to get all the feature of a terminal missing from Terminal.all

15:14:35 pjb s/all/app/

15:15:07 ToeTag lol

15:16:18 m_zr0 morbidly curious aboout that last one

15:16:43 ToeTag m_zr0, I don't know if you covered this already but CLISP, when started from the commandline has an interactive interpreter with, I think a GNU readline wrapper so you get tab completion of function names

15:16:44 m_zr0 and it's named cathode T_T

15:17:02 ToeTag and sbcl doesn't have that for me, that I know of - haven't read the whole manpage yet

15:17:40 m_zr0 it does! and it rocks. the issue was when it made workflow confusing for a beginner (me) to follow without newline/indentation

15:17:51 m_zr0 all those crazy paren. hugs

15:17:55 m_zr0 touching me

15:18:01 m_zr0 * runs screaming, then comes back

15:20:20 Petit_Dejeuner I'm trying to use ningle/clack to upload media files to a server, but I can only return a string at the end of my route without getting an SB-KERNEL:CASE-FAILURE. It's expecting a vector of unsigned bytes, a list, a pathname, or nil. If I try to coerce a binary file into a string, it doesn't work because not all the bytes are valid characters. Does anyone have an idea what I might be doing wrong or where I should lo

15:20:21 Petit_Dejeuner ok?

15:21:08 remi`bd I don’t know ningle/clack, but there’s usually special handlers to send files

15:21:10 pjb Didn't specify the right encoding, or else it's really binary data.

15:29:45 m_zr0 btw thanks for the help guys, working on all of this now

15:30:40 jackdaniel m_zr0: there is also channel #clnoobs

15:31:58 m_zr0 gotcha

15:32:37 m_zr0 .....why is everyone in there in their underwear?

15:33:02 jackdaniel what?

15:33:04 m_zr0 well, when in rome

15:49:45 digiorgi I am having a problem with cl-who, the code is simple. http://paste.lisp.org/display/152273

15:50:20 digiorgi CL-WHO print two times "es" instead of one

15:54:06 digiorgi i see it like a bug

15:55:57 sigjuice_ what libraries are available for html parsing or screen scraping?

15:56:59 beach sigjuice_: Closure (the web browser) has an HTML parser in it. As I recall it is also `fault tolerant' in that it tries to patch bad HTML.

15:57:14 beach The parser might already be maintained as a separate library.

15:57:21 beach But I don't know the details.

15:57:33 dim closure-html is in quicklisp

15:57:43 beach There you go.

15:57:44 dim with plenty others

15:57:49 dim (ql:system-apropos "html")

15:57:53 Fade cl-mechanize is also in quicklisp, last I looked.

15:58:44 Fade which is a pretty decent web scraping lib.

16:00:27 sigjuice_ thanks for all the suggestions. I want to write a program to periodically fetch my cable modem's diagnostic web page to monitor signal levels and other parameters.

16:03:56 Fade sigjuice_ : is this your first forray into lisp?

16:06:31 sigjuice_ I have been playing with list on and off over the past few years.

16:06:42 sigjuice_ s,list,lisp,

16:06:45 Fade *nod*

16:07:32 ToeTag Before quicklisp, did people just have to individually download libraries from their respective sources

16:08:03 sigjuice_ ToeTag: there used to be a system called asdf-install

16:08:15 p_l and in-between asdf install and quicklisp there was clbuild

16:08:52 p_l ASDF-INstall was pretty limited and reliant on Cliki which had ... less than stellar uptime

16:10:07 sigjuice_ some Linux distributions used a thing called common-lisp-controller

16:12:45 Fade I used clbuild quite a bit. as I recall it was quite useful if you didn't already have sbcl installed.

16:12:56 Fade it would download and build it for you.

16:15:54 p_l * was heavy clbuild user

16:16:18 Fade it was also nice if you wanted to hack on any of the libraries it downloaded.

16:17:10 Fade quicklisp is better in just about every other conceivable way.

16:35:01 pjb What was nice when you had to collect the dependencies by hand, is that you knew what you put in your programs.

16:35:15 pjb Nowadays, you just ql:quickload, and it brings anything you don't care.

16:36:15 pjb Also, there are good libraries that aren't included in quicklisp (yet), because nobody asked it.

16:36:43 pjb (granted, a diminishing number).

16:36:58 guthur I'm using static-vectors package and trying to store some integers for use with a foreign lib, but i can't seem to either get the right type-specifier or something is going a little wrong; it is working fine with single floats

16:37:34 guthur when ever i try cffi:mem-aref (short) a index that should return 1 i get 256

16:37:58 guthur i'm using (unsigned-byte 16) as the type-specifer for the static-vector

16:38:23 kawaii-imouto ** NICK pyon

16:38:39 guthur sorrry signed-byte

16:45:04 |3b| should work, do you have some example code?

16:47:30 novemberist Anyone else here who uses ac-slime and can confirm that is doesn't work anymore after the latest slime update from melpa?

16:49:28 guthur |3b|: i'll post some, but what i am finding is that a vector that should be #(0 1 2 1 2 3) is #(0 256 1 256 2 ...)

16:54:05 guthur |3b|: http://paste.lisp.org/display/152275

16:55:04 guthur sbcl 1.2.4

16:55:15 guthur on linux

16:55:57 guthur oops meant to use signed-byte, it makes no difference though

16:59:09 digiorgi novemberist: i am using now slime-company

17:01:09 digiorgi any CL-WHO alternative?

17:01:14 novemberist digiorgi: never used company-mode before (is it better than auto-complete?), but i might give it a try until the issue with ac-slime is fixed

17:01:55 digiorgi novemberist: https://github.com/digiorgi/digiorgi-emacs.d

17:03:12 digiorgi novemberist: i found company faster, and ac better

17:04:29 novemberist digiorgi: thank you

17:08:06 |3b| guthur: s/mem-ref/mem-aref/

17:08:21 fe[nl]ix guthur: you must use mem-aref

17:08:22 |3b| guthur: or pass the offset in bytes instead of elements

17:08:28 fe[nl]ix hahaha

17:09:03 |3b| and don't access past the end of the array :p

17:11:10 guthur hehe,

17:11:44 guthur |3b|: cheers

17:33:29 beach

17:33:29 beach /part #lisp

17:47:54 antoszka digiorgi: http://www.cliki.net/html%20generator

17:47:59 novemberist digiorgi: i tried slime-company and it's really faster. but the docstring buffer doesn't really show the docstring but something else

17:48:08 antoszka digiorgi: for cl-who alternatives