freenode/#lisp - IRC Chatlog

13:27:52 lukego Maybe check-it? https://github.com/DalekBaldwin/check-it

13:35:52 _death fiveam also has some random testing support

13:37:09 lukego check-it code looks really nice to me at first glance. heavy on a pattern matching library called OPTIMA that looks nifty too

13:37:26 lukego seems like the generators can be used independently of the testing framework (but I'll probably want that later)

13:47:05 lukego hm but being part of an automated test framework there maybe aren't the right knobs for generators. Seems like the random sampling takes a uniform value up to a size limit, whereas I'd probably like some other random distributions e.g. log transform to bias towards small values while still permitting bigger ones. maybe can put that into the framework. reads...

13:49:12 jmercouris OPTIMA is not nifty, not in my opinion

13:49:13 jmercouris I don't like pattern matching libraries

13:49:17 jmercouris cryptic syntax

13:49:25 jmercouris it is not so verbose to write (equal x y) is it?

13:49:52 jackdaniel that's not what pattern matching is for

13:50:01 jmercouris that's what is often used for

13:51:02 jackdaniel by whom? ,)

13:51:06 lukego check-it source is really easy to read with the patterns, for me at least who's worked in patterny languages sometimes in the past (otherwise surely cognitive overhead)

13:52:24 lukego guess it's also one of those things where if you're doing pattern matching you'll tend to choose basic data structures to suit

14:50:10 Josh_2 Hi

14:51:18 Josh_2 Wheres the best place to look for examples of the MOP in use?

14:56:56 Gnuxie[m] what do you want to do with it?

14:57:16 lukego I feel dirty when I call SB-INT:COLLECT. Is there an obvious way to launder that e.g. via alexandria or something?

14:57:52 phoe lukego: uiop:while-collecting, unless you're allergic to uiop

14:58:26 lukego thanks!

15:01:02 jackdaniel uiop:while-collecting has a bug, I don't recall what exactly

15:01:17 jackdaniel but it goes belly up for some inputs

15:01:36 Gnuxie[m] Josh_2: Postmodern/Mito/json-mop have readable examples for slot classes and other basic use cases (but honestly h*ck ORM)

15:02:27 jackdaniel lukego: collect originates from cmucl, here is one of many utility copies: https://github.com/McCLIM/McCLIM/blob/master/Core/clim-basic/utils.lisp#L716

15:02:59 phoe the issue is that I don't know a commonly used library that includes this macro

15:03:07 phoe flip214: maybe alexandria2 would consider adding it?

15:03:07 lukego I'm actually fine with being non-portable it just feels icky to reference SB-INT in source code because it looks "internaley." Back when I used CMUCL I was more than happy to use EXT:COLLECT :)

15:03:30 phoe s/whole post/maybe you would consider adding it to alexandria2?/

15:03:36 jackdaniel lukego: just copy it to your utilities file and forget about the issue

15:03:41 lukego Or otherwise how would you convert alexandria's MAP-COMBINATIONS into a COMBINATIONS function that accumulates into a list?

15:04:03 Josh_2 Gnuxie[m]: I'm not sure, I need to get a gist of what I can do with it before I can understand how I can apply it to current projects

15:04:07 jackdaniel the library "collectors" has somewhat similar equivalent

15:04:16 Josh_2 I will check what you have suggested, thanks Gnuxie[m]

15:05:14 lukego jackdaniel: yeah fair enough. I did the next best/worst thing of updating my defpackage with (import-from #:sb-int #:collect)

15:05:25 jackdaniel that works too, sure

15:06:31 lukego I'm curious though how other people usually do "collecting" with "mapping" functions though? I mean for data that's not supported by loop accumulating keywords and doesn't provide a reduce-like accumulating version

15:06:59 jackdaniel ACTION uses ext:collect on ecl and climi::collect in mcclim :)

15:07:03 phoe a LAMBDA over PUSH tends to save the day

15:07:28 jackdaniel don't forget about nreverse

15:07:59 phoe oh right, (let ((x '())) ... (lambda (y) (push y x)) ... (nreverse x)))

15:08:25 lukego ok yeah that's decent too but I'll stick with collect

15:08:53 lukego I have missed this about Lisp, curating one's own personal programming style :)

15:12:15 beach Don't take that possibility too far.

15:13:05 phoe "with great expressiveness comes great possibilty of making Lisp as readable as poorly golfed Perl"

15:16:36 contrapunctus lol

15:17:22 lukego shoe changing foot quickly :-) thought I was clever doing (collect ((comb)) (alexandria:map-combations #'comb list)) ...) but of course COMB is a not a function...

15:18:16 lukego not immediately obvious to me that COLLECT needs to use MACROLET instead of LABELS under the hood either...

15:19:08 phoe I think that FLET is enough though

15:19:36 phoe at least my lambda example up there has a single anonymous function for collecting stuff

15:19:47 contrapunctus beach: I had a look at the 'ancient CL version' of OpenMusic on GitHub...it seems to be GPL, whereas I prefer to Unlicense my work. 😔 I guess I'll look into wrapping the RtMIDI C API, which seems like the next most pragmatic route.

15:20:59 phoe ancient? https://github.com/openmusic-project/openmusic has a release from this year

15:21:03 phoe but that is GPL as well

15:21:10 jackdaniel the macro collect uses macrolet to allow having any collecting method, not only collect list

15:23:56 lukego collectors library looks nice but code doesn't seem to match implementation. (collecting (map-combinations #'collect list)) should do what I want but the macro doesn't parse those args

15:31:04 contrapunctus phoe: oh, I confused it with CommonMusic 🤦‍♀️

15:34:14 lukego so my ultimate compromise for now is to take the nice COLLECTING interface from the COLLECTORS documentation, which isn't really supported by that library as installed from quicklisp, and port it onto the top of SB-INT:COLLECT. good enough for now. https://gist.github.com/lukego/1714405ba893250812bf522a1e4f1b97

15:39:34 lukego then you can (defun permutations (list) (collecting (map-permutations #'collect list))) which is a darling I haven't murdered yet :)

15:51:04 lukego So next naive question perhaps leading to a hot library tip: how would you compute the list of all subsets of the set LIST?

15:53:00 phoe wait a second

15:53:05 phoe LIST === (OR CONS NULL)

15:53:10 phoe that's when it comes to types

15:53:23 phoe what exactly do you want to do?

15:53:26 jackdaniel that's not a problem, the resulting list will have a nil as one of subsets

15:54:10 lukego (subsets (1 2 3)) => (() (1) (2) (3) (1 2) (1 3) (2 3) (1 2 3))

15:54:56 lukego I find myself writing a recursive function but it feels like this should exist and that makes me wonder if I'm missing another alexandria/uiop sort of library

15:55:08 lukego or simply doing an apropos search on the wrong keyword..

15:55:51 lukego ACTION consults that awesome-lisp page again..

15:57:50 jackdaniel lukego: I have something

15:57:53 jackdaniel hang on

15:58:55 phoe (let ((x '()) (list '(1 2 3 4 5))) (loop for i from 0 below (length list) do (alexandria:map-combinations (lambda (y) (push y x)) list :length i)) x)

15:59:05 phoe could also use DOTIMES instead of LOOP

15:59:35 contrapunctus Doesn't loop let you say 'with x'?

15:59:42 phoe oh right, it does

15:59:59 jackdaniel what phoe wrote is ~same what I had in mind

16:00:50 pve lukego: is that the powerset?

16:01:02 pve https://rosettacode.org/wiki/Power_set#Common_Lisp

16:01:08 phoe pve: it is

16:01:11 lukego phoe: oh nice. I can combine that with my little NIH macro and have (collecting (dotimes (i (length list)) (map-combinations #'collect list :length i))) which is just fine

16:01:58 lukego thank you pve! I lacked that term and it's the right one

16:03:19 pve lukego: np

16:03:33 lukego I like this little collecting/collect pair from COLLECTORS interface because I'm not sure that I've ever actually had multiple collectors with COLLECT nor given a collector a useful name

16:03:46 lukego Jolly nice lisping with you all. nice weekend!

16:11:17 dlowe Seems like you could make a list type if you could do (deftype my-list (or (cons my-type my-list) null))

16:11:26 dlowe *a nicer list type

16:16:21 Bike recursive types introduce complications into the system that i, at least, wouldn't know how to solve. for example under the usual recursive definition of subtype, #1=(cons t #1#), which i would understand as the type of infinite circular lists, is a subtype of the bottom type NIL

16:46:57 frgo Hi all: question: I need to copy a large (2GB) in 255 byte chunks into foreign memory using CFFI. Would I use something like https://github.com/jetmonk/waaf-cffi for this?

16:47:18 frgo s/large /large file /

16:48:06 frgo Or just cffi:with-foreign-pointer and loop ?

17:01:41 phoe maybe try mmapping it instead

17:49:57 frgo phoe: Yeah. But but but ... How to get that into foreign memory "easily"...

18:03:46 phoe frgo: mmap it and memcpy?

18:04:51 frgo phoe: Hmpf - yup. You are right - it's the simple things that sometimes are hard to find ... Thanks!

18:39:38 Josh_2 Do we have a 'seen' command in here?

18:49:28 aeth I think minion has a seen command? iirc

18:49:42 aeth Josh_2: I usually just use this though: https://irclog.tymoon.eu/freenode/%23lisp

18:50:17 aeth you can search "written by" and keep expanding the time interval until the person you are looking for has spoken, but keep in mind that it's rate limited to a very noticable length, maybe 5-10 seconds... https://irclog.tymoon.eu/search?channel=freenode%2F%23lisp

18:50:49 aeth There's no all-channel search, though, so a bot has the advantage that it might have seen someone on another channel who hasn't spoken here recently.

19:07:25 Josh_2 pjb hasn't been around for a while

19:07:29 Josh_2 or did he change his name?

19:21:14 aeth pjb was online recently, but hasn't said anything

19:54:36 Josh_2 aeth: okay sweet

20:53:18 seok how do I call a function with variables in a list?

20:56:02 ebrasca seok: Can you explain more what do you whant?

20:56:37 jackdaniel (apply function arg-list)

20:56:58 Josh_2 Did Ravenpack fill that junior position in Marbella?

21:03:55 Gnuxie[m] Josh_2: pjb got banned i think

21:10:23 jackdaniel he got unbanned few days later

21:15:02 Josh_2 banned in here??

21:15:24 younder not for the moment

21:16:29 younder Truth be told I write drone software these days. Hardly any Lisp. I just hang around for old times sake.

21:16:40 Josh_2 write drone software in Lisp :P

21:18:22 mercourisj anyone know of any Lisp applications for photo management?

21:22:09 Josh_2 i'm looking back at some of the code I wrote over a year and trying to fix it up, It's dreadful oof

21:22:42 younder The easiest way is use a shell language interface like the one provided by imagemagic. Shell out to that and it should provide what you need.

21:22:48 mercourisj that's good news Josh_2

21:22:50 mercourisj that means you are growing

21:23:13 mercourisj I'm talking about photo library management, browsing, viewing, sorting, etc

21:24:11 younder ACTION confesses to mosty use imagemagic from PHP..

21:25:20 mercourisj maybe a good candidate for a CLIM application

21:26:04 younder Image processing is a lot of work.. Hope you have a lot of friends ;)

21:26:36 younder A frontend to OpenCV might be a start.

21:27:44 torbo What's the best way to parse a string representation of a rational number (i.e. "3/4") into a rational number in CL?

21:28:51 younder read

21:29:02 Bike easiest is READ but it'll do bad things on malformed input. you could use POSITION to get the solidus, and then PARSE-INTEGER both sides.

21:30:37 torbo Great, thank you both!

21:43:42 jasom there is also the library parse-number

21:46:12 younder Bike is a PhD. He writes robust code for industrial applications. I am guessing turbo isn't there yet.

21:46:23 Bike what the hell?

21:48:33 younder Bike a was a peripheral developer on you compiler

21:48:57 younder ^ O

21:49:04 younder ^ I

21:56:02 aeth here? I remember pjb being banned from #scheme for a short while for only being there to promote CL... that might be what you're thinking of.

21:58:38 Bike as jackdaniel said, already, less than an hour ago, he was banned from here for a few days

21:58:38 scymtym i think people are referring to this: https://irclog.tymoon.eu/freenode/%23lisp?around=1592014365#1592014365

22:04:36 younder I don't know by I have not been here. Left andforgotten and mostly irrelevant. If I say someting that sounds like misinformation, please let me know.

22:20:23 Josh_2 seems like a petty ban

22:20:56 Josh_2 This community is pretty small, last thing needed is isolating individuals over petty stuff like being abrasive

22:22:30 White_Flame there's a difference between being abrasive and being intentionally obtuse, screaming that everybody's correct and applicable answers are absolutely wrong in expletive laden screeds

22:23:28 Josh_2 Hey, that's just pjb xD

22:24:05 Josh_2 I started here when I was basically 100% new to programming and I got that impression from him pretty early, I still appreciate the walls of text to this day

22:35:38 younder White_Flame, sounds like a bot

22:52:57 Lord_of_Life_ ** NICK Lord_of_Life

23:07:53 dlowe It's *more* important in a small community to boot people who spread bullshit.

23:23:08 stylewarning dlowe: CL:BULLSHIT isn’t a part of ANSI Common Lisp. *DEFAULT-BULLSHIT-DEFAULTS* may have bullshit or other components of bullshit, though.

23:23:23 aeth pjb gives particularly bad advice on IRC when the question is about the type system ime... His advice is consistently to never check types and just let things fail at the point where they fail, but ime that just leads to a random NIL showing up 20 functions away from where it was generated and a waste of a day.

23:23:49 torbo jasom: Thanks!

23:29:52 johntalent How would it be possible to use ABCL with The Godot Game Engine?

23:30:29 aeth ABCL is Java-based and not very popular, at least on IRC, so you would probably have better odds with ECL, which is designed to be embedded in C applications.

23:37:46 aeth johntalent: It looks like all other (not GDScript, VisualScript, or C#) languages are supposed to interface through the GDNative C++ bindings. https://docs.godotengine.org/en/stable/getting_started/step_by_step/scripting.html#gdnative-c

23:38:08 aeth ECL or even Clasp (the latter is a work-in-progress) seems like it would be a lot less painful than trying to bring in the entirety of the Java environment

23:39:56 aeth johntalent: I found one result: https://github.com/borodust/bodge-godot

23:55:56 borodust aeth johntalent: i guess i need to add a warning there or smth - those bindings are highly experimental - i never used them it was just an experiment to see how hard it would be to generate bindings for godot

23:56:17 borodust it was easy to, but i didn't go any further, so those might actually not work ;p

0:03:57 borodust once i'm done with c++ support in :claw i might get back and actually try them out :)

0:04:53 borodust hopefully, that doesn't mean never, cuz i'm 95% done implementing it and only need to finish the other 95%

0:07:11 johntalent aeth: wow.

0:07:20 johntalent aeth: looks good thanks. somewhat recent too.

0:08:35 johntalent thanks borodust

0:08:38 johntalent thank you aeth.

0:10:13 borodust i hope that wasn't about my bindings to godot api...

0:10:51 borodust but i've added a warning so it should be alright now :/