<daviid>this name thingis fun and pretty booring at the same time :) <daviid>paroneayea: you're pretty good for finding names as well, help! :) <ijp>this is why you think of the name first, and then write the project to match the name <daviid>ijp: that's exactly why I submited early <ijp>I'd make the jump from introspection->inspection, and then pick a detective name, like zenigata, or poirot <daviid>paroneayea: not being an english native I need help to understand <daviid>an owl? roraring is what the lin does right? <daviid>paroneayea: would it stand for an acronym as well or just a 'joke' [words game] <paroneayea>daviid: just a joke/pun because of "GIR" (so "GIRowl") <ijp>did anyone use "GIRl" for GIR Library yet? <ijp>someone must be using it....dnf searches <daviid>ijp: I don't think so, let me find the page listing existing bindings <daviid>I must say I liked Gbank, is it because I picked it up :) or someone else here like it s well ? <daviid>so far gobi, swivel, girowl, poirot [or another detective maybe], girl <daviid>you see in the list we had guile-gir, abandonned, and not very sexy but ... <daviid>I'll think about and submit these, we'll see! thanks all <zv>what is the best guide to scheme (guile) macros <zv>I feel like doing something like "define a bunch of functions named fun-1 through fun-10" are fairly easy using CL-style macros but the documentation doesn't really suggest HOW that would be done (in the general case) using `syntax-case` <zv>wuw, thank you daviid ! <zv>many questions remain unanswered <zv>and I'm not demanding a SPECIFIC answer, just the building blocks here <zv>1. How to dynamically generate GOOPS fields/methods <zv>e.g (make-fields (thing #:virtual) (thing1 (#:accessor thing1) class), just as an example <zv>it was a great guide however! <zv>the core of syntax->datum (and vice versa) is what I'm really curious about <zv>is it best to just try to adapt to racket's documentation? <zv>Does anyone here know how to use a macro do define a list of functions? ***holomorph is now known as Guest2055
***holomorp1 is now known as holomorph
<galex-713>zv: when you say a set of functions, you mean like a potentially large set of similar functions with only part of their names changing? <galex-713>like (define (((pre post) fix) string fix) (apply concatenate (if (equal funname 'prefix) (list fix string) (list string fix))))? <galex-713>davexunit: but isn’t there anyway to support this through my computing softwarely calculating what’s not in that release? <zv>so galex, I ended up just doing it with eval + quasiquoting :( <zv>anyway galex-713 -- there is hope! <galex-713>Like to do ugly stuff like “(define se when)” <galex-713>ecraven: I see there are stuff like “macro-name” and “macro-transformer” but I can’t understand what does they return exactly and how to use them (makes syntax error anyway) <ecraven>but that is unhygienic, remember that <galex-713>what’s unhygienic yet? I heard it is like dynamic scope but for macros <ecraven>hygiene is related to capturing or introducing new identifiers <galex-713>well I’d like to have two lists like (when until case) (quand jusqu’à cas) and redefine the first symbols macros to the seconds <ecraven>someone more knowledgeable about guile needs to answer this, I don't know :) there should be something like syntax aliases <galex-713>But anyway, is it *possible* to programatically define macros? <ecraven>well, a macro can expand to the definition of other macros <ecraven>(define-macro (def a b) `(define a b)) <ecraven>rather: (define-macro (def a b) `(define ,a ,b)) <galex-713>I knew that, but can you define macros from inside a loop for example? <ecraven>can you show me an example of what you want? <ecraven>local macros can be achieved by let-syntax and friends <galex-713>well, as I said: programatically define a set of macros, like (define-mymacros (quando fino-a casi) (when until case)) <galex-713>so that I loop for each car of the lists and define the first to the second <ecraven>yes, *if* you find a way to define one of these, you can easily write a macro to define them all <galex-713>Yet I have to find something that a working (define-syntax se if) x) <ecraven>yes, if you have that, define-mymacros is easy <ecraven>just expand to (begin (define-my-syntax-alias quando when) (define-my-syntax-alias fino-a until) ...) <galex-713>is all the binding/transformer stuff explanation inside guile manual quite understandable or yet complex to understand? <galex-713>By curiosity, where did you learn macros, hygiene, etc.? <galex-713>you mean like the ones cited in the /topic of #scheme? or even more? <ecraven>well, read everything you can about all the interesting programming languages... Forth, Smalltalk, APL, Prolog, Haskell, things like that <ecraven>common lisp.. even the old interlisp and lisp machine manuals are very interesting <galex-713>oh I have to learn about lisp machines one day <galex-713>Never went deeper than the wikipedia article in reguard to forth and smalltalk, and never heard of APL <ecraven>I haven't had the fortune to actually experience one, but looking at youtube videos and reading the manuals they certainly seem very interesting <ecraven>APL is interesting because it's a vector language.. J would work fine here too, and many others I never heard of :) <ecraven>specifically for macros, just google for syntax-rules or read the right chapters in the /topic of #scheme <ecraven>some people like Let over Lambda, a book about common lisp (indirectly applicable to Scheme) <ecraven>galex-713: the relevant keywords are probably: lisp machine genera symbolics <civodul>"The R7RS restored string-set! and string-fill!" argh <davexunit>I don't get a very good impression from the folks making these scheme standards <civodul>i don't have a good impression of R7, though i haven't read closely <peterbrett_work>I still pretty much stick to R5RS — it has everything I'm interested in in it <civodul>it would be nice if many of us could meet there, and talk about our Guile things <galex-713>Although last year there were only something at FOSDEM I heard <davexunit>galex-713: re: sly: no, it's not possible to use a lower GLSL version because it really reduces the features I can use. <galex-713>davexunit: no I mean is mandatorly that GLSL version hardware or is it possible to have it software-implemented on top of what hardware provides? <galex-713>*is it mandatory that GLSL version is fully provided by hardware or […] <davexunit>galex-713: the GPU drivers need to support that version <galex-713>why? can’t a layer be added between the GPU and the GLSL? <davexunit>the OpenGL specification says that GLSL source code must be given to the driver <galex-713>Annnnnd cannot a fake-glsl-library provide GLSL 1.3 while giving 1.2 to the driver? <davexunit>because it won't support the features necessary <galex-713>can’t glsl 1.3 stuff that can’t be done with 1.2 be done with software rendering? <davexunit>maybe some compatibility hack could be done, but I don't know any prior art and I imagine it would be enormously difficult. <davexunit>bottom line is that your GPU needs to support the versions of GLSL introduced with OpenGL 3.2 <davexunit>which I think is actually 1.5, which I plan to bump the shaders up to using. <galex-713>so if the ThinkpadX60 only supports 1.2 GLSL I have to find something else than Sly, maybe inspiring of sly you mean? <davexunit>perhaps the rendering engine could support a degraded mode with less features, but I have no plans to write something like that now. <galex-713>I suppose it should be non-trivial to make a game engine with 1.2 if you did with 1.3 and are planning of doing with 1.5 right? <davexunit>you'd just have less features available to you <peterbrett_work>galex-713: Where are you from where 50 EUR / day is expensive for a conference o.O <galex-713>Currently the fee is what I pay monthly to rent my appartment :p <galex-713>so yes, pay per day for a meeting what I’m used to pay per month for basic living is a bit expensive to me ^^ <peterbrett_work>galex-713: My apartment is £875 / month and most people think that is a bargain! o.O <peterbrett_work>It is a nice apartment but it's hardly luxurious. Around here (Edinburgh) you would not be able to rent a bed in a homeless shelter for EUR 50 / month. <civodul>galex-713: that includes food, but yeah, you can ask for reimbursement when you show up <civodul>the idea is to make sure they don't buy stuff for people who don't show up, i think <galex-713>peterbrett_work: well I maybe lied a bit since I divided that by the number of people I live with <galex-713>And removed from that social prestations before <galex-713>civodul: Yes I understood that, but I was wondering how can food be so much expensive? does that also include the right to do something at Inria? <galex-713>(and yet there are lot of void old unused building everywhere, including some dozen of millions of free and void appartment everywhere) <civodul>galex-713: i don't know, i'm not organizing it :-) <galex-713>cgrothoff still didn’t answer on #gnunet he must be buzy ^^ and my mail system is increasingly confusing me <galex-713>(or maybe is it the contrast GNUnet+secushare guys built in my head along with the strong bias against SMTP that makes me feel that) <galex-713>peterbrett_work: you know I managed to pay only 100€ for the whole 32C3 and 50€ for the 31C3 (including hosting, food, travel, tickets, etc.) <galex-713>so yes it is possible to find that expensive for a conference :) <galex-713>yes, France is near Germany without any water in between so that makes travel easy ^^ <galex-713>(especially carpooling+autostop possible \\o/) <galex-713>peterbrett_work: is the UK-France tunnel so expensive? (don’t know anything about it) <galex-713>Ok I asked, it actually really all for the food, because grothoff is going to reserve catering <galex-713>is there any binding between guile and csound? or guile and any sound-synthesis stuff? <fantazo_>how serious is the guile project with "extension language for gnu"? I mean judging from the past it's not that serious. people always loved to embrace something different, like TCL, Perl, Python, etc. <galex-713>davexunit: and except sly, is there any functional graphical stuff? <davexunit>people have done other graphical stuff, but I haven't tried them <fantazo_>galex-713, normally people run off and create something only that it's shortly after forgotten and unmaintained, welcome to guile :-) <galex-713>fantazo_: guile is the primary package to make GNU package extendable, I think GNU coding standard recommand to use Guile as extension system, although Guile could perfetely implement TCL, Python and Perl (by order of ease, by guess), already does for scheme, elisp and javascript, and will soon for Lua <galex-713>davexunit: so sad :( do thing works that way too in other ecosystem like python, perl, etc.? <davexunit>yes, most libraries people make are unmaintained <davexunit>guile just has less people writing libraries <fantazo_>galex-713, well that's what the package reads. still theory and reality are two worlds. <galex-713>fantazo_: what do you mean by “what the package reads”? <fantazo_>galex-713, that's what the gnu project suggest, not that it's lived reality, even inside "the gnu project". as it's all a bunch of different people who want to have "fun" in their free time and holding to a common thing isn't part of that. <galex-713>quite most known GNU packages *practically* integrates guile whenever possible, scheme/elisp&js *already works* (I can run them now in my terminal with guile) and I think there’s a branch for lua inside the git, although if you look at the manual the API is very well documented and there are a lot of bindings with a lot of things, although if you document on scheme it’s a really <galex-713>neat language and guile makes extension as simple as possible, so with little effort you can make a programme extensible with several languages, and that gives a very cool work/benefit ratio ;) <galex-713>ijp: well I tested code and it worked, that’s enough “works” for me <galex-713>fantazo_: I was just answering your remarks on the seriousness of guile :) besides I can understand why someone could strategically want to run a primary implementation like python or LuaJIT to extend their programs, but these are up to devs choice and there are enough people working on guile to consider it useful ^^ <fantazo_>galex-713, you misunderstood me. I didn't say anything about the seriousness of the guile community or the guile project. but more of the seriousness of the "gnu project" of using guile. like gnome for example, the last time I checked they use something differently for their gui hacking needs. but maybe I'm off, I haven't looked at gnome for years. <ijp>gnome has no interest in guile <galex-713>fantazo_: gnome is completely away of the gnu project x) <galex-713>I think they had gone almost open-source people <galex-713>loving cloud stuff, minitel 2.0 and other terrible stuff like that <fantazo_>galex-713, well then the wikipedia article is wrong about gnome then. <davexunit>yeah, GNOME has no real relation to GNU at this point. <galex-713>fantazo_: historically GNOME come from GNU, so theorically it’s inside GNU project, practically it’s two completely separate communities with different ideas <dsmith>Kind of like the relationshp betwen Java and Javascript. <galex-713>And I can understand that since that’s two different system programming layer: sociology of programming teach us why these people are more close to web devs than to C/lisp hackers like you can easily find into GNU <fantazo_>galex-713, how are those two different layers? <galex-713>fantazo_: so yeah, *most* of “GNU Project” *has* a point into guile because GNU Project history is quite well linked to lisp and extensibility it can provides, notably because of GNU Emacs, and afaik GNU Guile was originally an attempt to bring GNU Emacs power and ideas to other programs <fantazo_>C, ok. but lisp? comme on lisp is considered to be a "high level" language. <galex-713>fantazo_: currently GNOME3 do things with javascript, CSS, a little C, lot of OOP, and a looooot of supercomplex abstraction like you can find into browsers <galex-713>While when you use stuff like emacs, lisp, guix, etc. even if you can do quite abstract things you’re still pretty close to the internal working of computer, I mean implementing lisp is quite simple&easy while a web browser is something really huge and completely different set of people do these things <fantazo_>so that their apps use 90% of the cpu time and resources and are constantly swapping. <galex-713>fantazo_: probably because GNOME people are used to huge 3D resources, video acceleration, HD-everywhere because of the programming milieu they work in, then more probably most of them got more powerful computer, then they less care of less powerful computers <fantazo_>galex-713, one kind understands what they are doing and the other doesn't as one kind get's crazy when trying to implement something which is only partially understood and the other kind can deal with not understandable code which joe (tm) wrote while drunk. <galex-713>While if you backlog you can observe some Guile libraries can’t even work on my old computer, which is actually the same as rms (but tablet version and w/o bluetooth) and the first supported by GNU Libreboot <fantazo_>galex-713, not QTs? I mean that's the whole reason for KDE, then Gnome. In my opinion a stupid move. atleast two communities who are basically re-developing the same thing over and over again. <fantazo_>but on the other hand there are people with different needs. <galex-713>fantazo_: yeah Qt did something technically bad but socially necessary I think <galex-713>I mean GNOME/KDE/XFCE/etc. are necessary for most of people coming to GNU/Linux <fantazo_>and normally those needs are not mine. as I really find the whole mouse pushing idea an unwise one. <fantazo_>atleast there were some people who are not that demented of re-developing dead ideas and making atleast something minimally decent like tiling wms. <galex-713>Maybe one day we’ll finally succeed to make really clean, accessible, semantic, functional and modular graphical interfaces, but for now Qt/GTK+ are good to keep people on GNU/Linux and good enough to compete with Apple&Microsoft <fantazo_>when I read about plan9, they had some interesting ideas. like throwing out the whole term thingy. which is in my opinion a good move, as I'm not a big fan of terms, they are full of legacy which was topnotch 50 years ago. <galex-713>fantazo_: but there is muuuuch much much legacy to remove, and Plan9 only throwed ideas, not removed all the legacy <galex-713>Take Unicode/UTF8/ascii-based stuff, or qwerty keyboard, that still inside <galex-713>Although I didn’t heard of Plan9 as being a dream of accessibility <fantazo_>galex-713, sure. I'm talking on the level of only reading about it shallowly, so it must look like gold to me. <galex-713>And that is a big problem while at least today term provide an universal accessible interface until we get something really better <fantazo_>I think something like a term is a good thing, entering text and this is then magically handled. <galex-713>good enough until we have something really good :) <fantazo_>but with a more graphical interface. emacs has a lot of moments where I think it goes in the right direction. <fantazo_>like when you open a svg file and it renders it already, that's cool. <galex-713>Yet I think with functional graphic system, more semantic layer, and a whole OS like that, emacs points the direction :) <galex-713>fantazo_: yeah, except svg is hideous xml ^^ <fantazo_>galex-713, sure. but that's not a "xml is bad discussion", only if you want to make it one ;-) <galex-713>my conception is high level interpreted on-purpose languages with formal ontologies and which are translated to lower level interpreted language, until the hardware level (matricial pictures, raw sound, network packet, etc.) <galex-713>fantazo_: I think everyone agree sxml is good enough not to go into this discussion ^^ <galex-713>Even if personnally I go further and think even sexp aren’t good for other stuff than rendering generic code/data and we should restart doing binary stuff but helping ourselves with formal ontologies and cryptography <fantazo_>well, when you mean "binary structured data" then there are some specs for that. it basically doesn't matter as what you render a tree, what a s-expr basically is. <galex-713>oh you know things on that? :D I’d like links <galex-713>(I know W3C also tried to serialize XML but more-or-less failed) <galex-713>oh, it is really interesting, even if I was more wondering about contextual microformats than unique binary json <fantazo_>binary json, what's the difference to an s-expr? <galex-713>a serialized/binary s-exp would be as well the same <galex-713>I was speaking of something else, less universal: imagine a multitude of microformats who do represent only one thing, but do it well, imagine them contextually combined one inside the other, and imagine each one is identified by the unique hash of its ontology, that’s my idea <fantazo_>so you could substitute both, the only thing is again as always "personal preferences". <fantazo_>well, you haven't yet carefully looked at one of those. I know atleast superficially from cbor, that you can "extend" the types of the format. <fantazo_>so it's basically a keystore with some base types and then those which you want to add on will. <fantazo_>well, whatever. personal taste discussions, they don't make any sense. my life is too short for this. <galex-713>davexunit: is opengl really useful for 3D? or can SDL2 without it be sufficient? <davexunit>galex-713: without opengl you are greatly limited <davexunit>hardware accelerated graphics is a requirement for any serious graphical applications <galex-713>davexunit: beyond opengl API, do you know where to find resources on the extend the absence of opengl is limitating and what can it enables to do? <suppi>doesn't SDL2 have some hardware accelerated graphics? albeit limited in features <davexunit>if you're doing hardware accelerated graphics on GNU/Linux, you are using OpenGL <suppi>oh, they said sdl2 without it. sorry. I misunderstood <galex-713>what’s the difference between opengl and vulkan? <davexunit>Vulkan is a brand new specification seeking to improve on OpenGL <davexunit>by producing an API that more closely models how actual GPUs work <davexunit>galex-713: because it's not backwards compatible at all. it's a completely new API <galex-713>and do that mean new hardware will have to be made with it or new drivers will have to be made with it? <galex-713>davexunit: can openGL be used in a really functional way or is the imperative paradigm bound into the way it works? <davexunit>but it's unlikely that old hardware will have vulkan drivers made for it <davexunit>galex-713: it's a big state machine, so it's imperative. <davexunit>ligbl is a C library. nothing functional about it. <galex-713>Do you think that could be easily made functional with a layer uppon it? <suppi>there have been attempts at wrapping opengl with a functional style <davexunit>galex-713: I wrote a functional interface to opengl <galex-713>Did nothing survived or nothing got standardized? <galex-713>davexunit: but purely 2d and using glsl 1.3 right? <suppi>well, successful attempts perhaps? <galex-713>davexunit: I thought sly was purely 3D, or did you made a complete layer below sly which includes 3d stuff? <galex-713>Oh, I thought sly were intended only for 2D games <davexunit>it's just what I have written the most code for <suppi>haskell related stuff though <galex-713>If attempts were made, haskell couldn’t not being one of them ^^ <davexunit>galex-713: one of the reasons for changing the name was to get rid of the "2d" part <galex-713>davexunit: oh I thought that was because Sly was a more attractive name <galex-713>suppi: racket’s graphic stuff first made me think about that (functional graphics) <suppi>so plenty of stuff now that I think about it! <galex-713>I liked the racket stuff but I was like “oh, it’s only 2d and probably not vectorial” <avoine>davexunit: one thing that I miss in sly is basic geometry like curves, circle, etc but I guess I could use cairo to draw into an sdl surface right? <suppi>davexunit, I saw your danmaku engine a few weeks ago, great stuff! :D <galex-713>I’d really like to test and try to do things with Sly but it seems I have to wait I can buy a new computer before :/ <davexunit>in general I would like a high-level interface for making geometry <davexunit>so you could slap together some combinators and make fun things <suppi>galex-713, me too, but I'd need to have linux first :( <galex-713>suppi: do you only have access to work/school computer? <avoine>davexunit: I would love to help you with that <suppi>galex-713, i have a computer <galex-713>(I mean why isn’t there a GNU running on it) <suppi>galex-713, i can, i have reasons <davexunit>avoine: the basic idea is that the triangle is the primitive polygon, so if you can construct triangles and have some procedures to combine triangles in simple ways, you build complex things. <galex-713>suppi: related to work/school I suppose? or maybe games? <davexunit>and then there'd be some procedure like geometry->mesh that would transform that geometry description into an optimal OpenGL mesh <suppi>galex-713, it's a macbook air. I don't have a lot of incentive to move to linux <galex-713>suppi: oh I see, anyway I already observed strange behavior with GNU dualboot under mac <galex-713>Do you plan to try anyway or do that on another computer? :) <davexunit>so there might be a minor issue or two to sort out <suppi>galex-713, my next computer will probably be linux. I've tried to run sly in the past but it wouldn't work on OS X <davexunit>I no longer use guile-sdl so the big blocking issue is gone <galex-713>suppi: maybe together with davexunit you could help make Sly work on Mac OS X, could be great <suppi>davexunit, last time we talked i discovered that the problem was with compatibility issues with OpenGL <galex-713>davexunit: also why not making guile-sdl2 a gnu package along with guile-sdl and make both two modules in one package? <suppi>in OS X you can only use Core or something, and you can't use Compatibility mode features <davexunit>suppi: OK so there's an SDL function for setting that <suppi>it's about the features of OpenGL that you use in sly that work on linux but won't work in OS X <galex-713>davexunit: and will it one day be released with Guile or as a GNU project or something? ^^ (just to know) <suppi>something about mixing Comp and Core mode <davexunit>I don't use OS X so I can't fix the problems <suppi>I can check the logs if you want, we discussed it once <davexunit>galex-713: no, probably won't ever be a GNU project <suppi>oh, I probably don't have them anymore... <galex-713>davexunit: do you think if I buy an X200T to try Sly on that would work or is it not worth? <davexunit>avoine: that's basically what I already have with make-cube <davexunit>avoine: which is fine, but it's not a combinator interface <davexunit>so that you could define a cube in terms of 4 squares, where each square is defined in terms of 2 triangles <galex-713>what’s the differences between slayer and sly? and why have they similar names? is that linked to some english word? <galex-713>and any major differences or did the two project just never meet together/heard about the other? <davexunit>slayer takes quite a different approach to things <davexunit>I made sly with the specific intent of exploring functional reactive programming and live coding <galex-713>oh ok, while not slayer doesn’t you mean? :/ <galex-713>davexunit: hey, I’m really getting excited at Sly, isn’t there really any way of testing it? like, dunno, running it into a VirtualBox VM maybe? <davexunit>galex-713: that wouldn't work because the vm will be limited by the underlying hardware in terms of graphics acceleration <galex-713>oh, and there is no free-software VM able to simulate a certain hardware? :/ <galex-713>davexunit: and no software-rendering library that could simulate GLSL 1.3 can exist? <davexunit>certainly nothing could be compatible with sdl2 and the opengl api <galex-713>davexunit: and no way to compile sly disabling the 1.3 features? <galex-713>like if I don’t care and just want to test could removing stuff work? <davexunit>galex-713: you can attempt to rewrite the shader files to use an older GLSL version <davexunit>the files declare the version of GLSL they use <davexunit>with some small tweaks they could work with an older version <galex-713>davexunit: by curiosity, is font rendering in graphical program is implemented by graphical library (like GTK or SDL), opengl or something else? <davexunit>galex-713: I currently use sdl2-ttf to render text <wingo>re: software rendering of glsl 1.3 <davexunit>which renders to an sdl "surface" with a software renderer <wingo>ACTION uses cairo to render text, even though it doesn't kern as nicely as i'd like <davexunit>I then take that surface and make an OpenGL texture from it <galex-713>ok, now the next question related to that: is there any widely used way of supporting metafont into graphical programs? I mean something else than dvi/pdfs <galex-713>wingo: what did you mean about mesa&llvmpipe? <galex-713>wingo: do that mean I or davexunit would have to do something with it in order to make 1.3 GLSL Sly work here? <wingo>in general to use glsl v 1.3 you need a system that supports glsl v1.3 <wingo>and the corresponding opengl <wingo>mesa+llvmpipe is one implementation of that system; as long as sly is a conformant opengl program there should be nothing sly has to do <wingo>it's rather the problem of the person running the program to configure their opengl impl <galex-713>wingo: so that makes possible to glsl1.3 programs to run even if my GPU doesn’t support it? <wingo>lol i didn't realize ice-9 binary-ports already exported get-string-n! <daviid>wingo: I sympathize with the difficulties your're facing trying to reorganize these modules, looks like running after a snake escaping every time you're about to catch it :) <daviid>heelo guilers! fun thuesday all... <zv>Hey guys, I'm having quite a time writing the constraint solver system from SICP -- is there any way for me to "reload" the buffer in geiser (repl) every time without getting complaints about generic-functions? <zv>galex-713: essentially <avoine>zv: you can try (setq warning-suppress-types nil)