<civodul>daviid: you need to be admin of the project and to log-in on Savannah <daviid>civodul: ah ok, I am a member but not an admin <daviid>rlb: you're a g-wrap admin, can you add myself as an admin maybe? <daviid>rotty1: ^^ you too coiuld add me I suppose <rotty1>sneek: later tell daviid are you tupi @ savannah? ***michel_mno_afk is now known as michel_mno
<amz3>how can I do something like alist-ref ? <tohoyn_>scheme@(guile-user)> (atan (tan -0.5+2.0i)) <tohoyn_>$1 = 1.070796326794897+1.9999999999999996i <tohoyn_>scheme@(guile-user)> (tan (atan -0.5+2.0i)) <tohoyn_>$2 = 0.1176470588235294+0.4705882352941177i <amz3>it's a poc for artanis website <amz3>if the guile-log fails to build on prolog things it's ok you can still `make build` and `make serve` <amz3>next parser will be for wikipedia I guess :) <stis>amz3: I plan to make it possible to decide in ./configure what to get. <stis>e.g. prolog or not cause it is heavy <amz3>IIRC a file is missing in git <stis>Oh is it in the earlier checkout you did or where I work at the head? <nalaginrut>davexunit: I've heard that you want to wrap markdown parser with FFI rather than write a new one in Scheme? ;-) <davexunit>nalaginrut: I want a pure Sheme one ultimately <davexunit>but I want to quickly wrap the hoedown library with the FFI <nalaginrut>alright, we may have a pure Scheme one later ;-P <nalaginrut>I spend all my hack power on async server core, or I can write a buggy one quickly <davexunit>pure scheme would be most flexible, but slower. <stis>is the parser the speed limiter? <nalaginrut>davexunit: come on, efficiency is not necessary for a static-page-generator <nalaginrut>if not too slow to generate a page in whole night... <davexunit>I sound like a broken record at this point, but I want to implement it using parser combinators where the input characters are represented as a SRFI-41 stream <davexunit>I just devote all of my time to guix so it never happens. <stis>do the parser combinators ontop of minikanren no? <nalaginrut>davexunit: amz3 has written a markdown parser with guile-log, maybe you want to take a look <davexunit>stis: I want 0 external dependencies and to do them in a monadic style <davexunit>I have a un-version controlled scheme file sitting on one of my computers that does it, more or less. <davexunit>I'm also not confident working in the relational paradigm yet <stis>minikanren is really small, you can include it in your sources and get 0 dependencies <davexunit>I don't think relational programming is necessary here. <davexunit>the implementation with monads is very nice. <nalaginrut>IMO, we can rewrite many thing in modern way and Guile's cool stuffs, so if any possible, I don't want to introduce any unmaintained lib too... <dmiles_afk>in some uses of guile-log are people not using the prolog syntax exension? <stis>well it gives you backtracking <stis>dmiles_afk: yes, I would say that people knows mostly scheme here and prefere kanren over prolog <davexunit>stis: is there a distinct advantage? my parser combinator implementation is LL, which I guess means that it doesn't handle left-recursive grammars. would using relational programming make it easy to write a GLL parser that can handle left-recursive grammars? <stis>davexunit: let us refhash this as f -> f g no? <dmiles_afk>opencog uses guile as their default scheme right? <stis>parsing f is like parsing m or parsing f then g <stis>that's inherent in the functional meaning of these parsers no? <davexunit>I'd like to stop the recursion of f if g can be parsed <davexunit>I've seen an implementation of GLL in racket but I haven't fully grokked it <stis>davexunit: why is it so important to have left recurrsiveness, can you examplify <davexunit>I haven't fully grokked it yet, honestly, and I need to refresh my memory about this stuff. <stis>If your mind is wired in thinking in parser combinators I don't see a left recursive feature as especially important. But maybe i'm wrong. <stis>I written a prolog parser and a python parser. I really didn't have any problems with left recursiveness. <davexunit>I have more to learn, which begs the question if I should be the one writing a parser combinator library ;) <stis>I actually postulate that you will be aböe to get a "left recursive free" parser spec for almost any language out there if you search. <davexunit>yeah, it could very well be not worth pursuing. <davexunit>maybe I should just dump my parser combinator module on guile-user and look for critiques <stis>davexunit: that's a good idea, also you are right about a 0 dependency for this framework is a good idea. <stis>I just wish that kanren was included in guile. <stis>I have made the original kanren sources work on guile. And iI think we have a kanren in guild-hall, just that I don't know if it is minikanren or full kanren <stis>guile-log pretty much sports a kanren interface as well <dmiles_afk>(that one is supposed to be full.. was the version you ported in CHICKEN scheme?) <stis>it was a long time ... let me check <stis>dmiles_afk: no I lost it in the last upgrade of my computer <stis>but I think I remember that it was the full version <dmiles_afk>i understand.. i have lost things that took me sooo long to create.. one such loss was a codemerge of EcoLisp and SWI-prolog <dmiles_afk>(so there was no datatype conversion when being called from one or the otehr) <dmiles_afk>secretly i need to understand why everyone only uses miniKanren <stis>dmiles_afk: I based my interface on the that kanren source code <dmiles_afk>hoping just now that most of this time you'd been thinking about kanren vars and such in guile-log you hadnt been religated to the miniKanren-only world.. but are mentally in the full kanren without saying <stis>well I think that full kanren and minikanren uses the same kind of variabling <stis>but yes I am living in the full kanren world <stis>I referred to minikanren because of the size of that codebase and perhaps also that it might be in pubklic domain <stis>not sure about those things. <stis>ok got to go CU hackers. <sneek>Welcome back daviid, you have 1 message. <sneek>daviid, rotty1 says: are you tupi @ savannah? <daviid>rotty1: yes i am, and speaking of it, I would like to change my savannah login for daviid, did try but didn't find how to do it :) <rotty1>daviid: I suspect your username is immutable <daviid>just asked on #savannah, but no answer <rotty1>daviid: I now made you an admin on g-wrap <rotty1>daviid: thank you for putting out a g-wrap release :-) <daviid>wc! i'm working on a guile-gnome release now, then guile-clutter <daviid>i'd like to split these on savannah as well, because they are really 2 different beasts <daviid>rotty1: are you still working on sbank? <daviid>rotty1: I did see variaous guile gir projects, 2 or 3, can't remember, I wonder what would be the best one to pick and start to work on, what is your idea? <rotty1>daviid: I don't know the current state of the "competitors" for sbank (my own spin at it). sbank is/was running quite nicely on Racket, but Guile's missing some basic functionality -- mostly the dynamic-link deficiency due to the libltdl brokenness recently discussed on guile-devel <daviid>rotty1: nothing could be done to overcome this [filename w/wo extension iirc] problem? is that why you stopped? [like I wouldn't be able to solve it either then] <rotty1>rotty1: (working on sbank) not ATM, but getting Guile up to speed wrt. (spells foreign) is still on my TODO list. The probably not-so-great thing about sbank from a Guile perspective is that it depends on quite a few libraries not part of Guile. <daviid>rotty1: just added and posted a news item for g-wrap, but it says 'A news manager of this project will have to review and approve the news.', are you the news manager ? [can't see anything about that] <daviid>can't access guile-gir: 'Gitorious.org is migrating all the repositories to Internet Archive. The repositories will soon be available for read-only access, with original clone URLs preserved.' but it's been ages they display this message and nothing is becoming available ... <daviid>rotty1: sbank was started before guile-gir, I think, but did you looked at it maybe? <rotty1>daviid: no, but I probably should <rotty1>(I've been having a mostly hack-less last two years, and am just now beginning to pick up speed again) <daviid>rotty1: I understand. well for the time being guile-git is unavailable anyway <rotty1>In theory, all that is required to get sbank running on Guile is implementing (spells foreign) and (spells gc), IIRC <daviid>Ok. I did not looked at sbank yet though, and from what you said earlier, if it depens on quite a few other lib not available in guile, maybe I would not be 'so much' interested :) don't know yet <daviid>right now i want to get guile-gnome 2.16.3 out, then guile-clutter 1.12.2, then I should try to work on more recent clutter version <daviid>rotty1: i'd like to split guile-gnome and guile-clutter at savannah, the savannah main page, git repo, homepage and documentation: what should I do, fill a new project request ? <sneek>I last saw wingo on Jun 24 at 08:05 pm UTC, saying: and how long prerelease will take, no idear. <rotty1>daviid: well these other libraries work on Guile, there just not neatly packaged <rotty1>(they are available as dorodango packages, but I really need to set up a dorodango repo, work on the documentation) <daviid>oh rotty1 by the way: guile-lib does not seem to have a pc file, would like to add me as a guile-lib member I could try to add this, it would be nice because of g-wrap, guile-gnome[clutter] dedpend on it and there is currently no way to check for the version number <daviid>rotty1: you mean an sbank [and dependencies] guild-hall package(s)? <daviid>rotty1: still no g-wrap news item, did you find out who the moderator? <daviid>we need to fix goops, if andy remains unavalaible for much longer [hope not!] <daviid>ACTION has always wanted to fix and maintain goops, oh well so much to do <daviid>rotty1: I know you're a goops pretty good user, are you also aware of its implementation? <rotty1>daviid: not really; it's been a long time, and I did not really look into goops itself, although I played around a bit with tinyclos <rotty1>daviid: I also get the message "only news manager can approve news". Probably some obscure permission setting... <rotty1>daviid: you're now a member of guile-lib <daviid>oh weird isn't it? #savannah doesn't answer any quizz either <daviid>rotty1: ah, many thanks! will work on a .pc file soon <civodul>daviid: BTW, congrats on the new G-Wrap release! <daviid>civodul: thank you! same for guix, amazing work! <daviid>civodul: do you know who's moderating 'news' entries, per project, on savannah? <civodul>sneek: later tell daviid news entries are moderated by each project's admins