<paroneayea>so, pretty dumb, but I'm having trouble figuring out how doing something like this would work in guile <davexunit>I haven't done too much work with pipes, but this involves using the pipe API <paroneayea>that was my guess, but I can't figure out how to tell the pipe "I put this data into it but I'm done now" <bipt>(force-output port) maybe? <paroneayea>(the second (display) was for me to verify it went past there) <davexunit>paroneayea: I can't put my finger on what's wrong, but I don't think OPEN_BOTH is correct <davexunit>off topic, but does anyone know a good way of creating a temp module to produce a fresh top-level environment to eval code in? <davexunit>but now, how to programatically evaluate something in the top-level... <bipt>there's also make-fresh-user-module if you want the default guile bindings <davexunit>paroneayea: I did this exact thing in december but the code is gone <paroneayea>davexunit: yeah I should learn to use the FFI. I'm not there yet ;) <bipt>and eval takes a required module argument <davexunit>I'm writing a little example to go with Haunt, a static site generator I'm working on <bipt>paroneayea, i was wondering about that. i tried close-output-port but that also closes the input side <davexunit>and I want to write a macro that generates sxml that shows the code in plain text, and the result of evaluating it below. <paroneayea>bipt: yeah python gives you separate things to access for stdin, stdout, stderr <davexunit>I'm too distracted right now, but I totally did this awhile ago and it made sense <paroneayea>I'm also confused as to how it doesn't seem that there's an *option* for getting stderr in open-pipe* ***karswell` is now known as karswell
***mario-go` is now known as mario-goulart
<davexunit>ArneBab: I added the crappy blog builder and an example site to the haunt repo. <zacts>hi davexunit, mark_weaver, civodul, et al... <zacts>davexunit: I think I noticed you on #lispgames <zacts>are you doing common lisp game related stuff, or just chatting about your guile 2d game lib? <davexunit>zacts: chatting about sly and with dto about his games, mostly <zacts>that spaceship 2d / fake 3dish game <zacts>I've been working on my math proofs <zacts>in preparation for SICP and other things <davexunit>the fights look like a more abstract version of afterburner ***michel_mno_afk is now known as michel_mno
<ArneBab_>sneek: later tell davexunit: cool! thank you! <sneek>Welcome back davexunit, you have 1 message. <sneek>davexunit, ArneBab_ says: cool! thank you! <ArneBab_>for haunt: since you have the example in a separate dir, could haunt build take an argument (filepath), so people could build multiple sites from the haunt sourcetree without adjusting GUILE_LOAD_PATH or installing haunt? <davexunit>ArneBab_: the pre-inst-env script sets things up for running haunt from the source tree <davexunit>all haunt commands accept a --config=foo.scm option as well, to change the config. <davexunit>I bet site building breaks if you aren't in the same directory as the config file, though <davexunit>I haven't determined the right behavior for that. <davexunit>I think it's reasonable to allow a config to be pulled from anywhere, but the site building happens relative to the current directory <ArneBab_>why do you use absolute paths for the site? <davexunit>I think you should treat 'haunt build' like 'make' <davexunit>run it from the root of your site's source tree, just like the Makefile <ArneBab_>in the main directory: scripts/haunt build --config=example/haunt.scm <ArneBab_>what works is: cd example/; GUILE_LOAD_PATH=../ ../scripts/haunt build <davexunit>because that's the directory your posts are in <davexunit>instead of tweaking the load path, use pre-inst-env <ArneBab_>“instant usage: cd example; ../pre-inst-env haunt build; firefox site/index.html” <davexunit>to fire up a guile web server with the appropriate document root <ArneBab_>I just asked myself how I could insert a haunt site into Freenet (though the absolute URLs might just work, I’m not sure) <healer>wingo, what is the current state of guildhall? <healer>is it suitable for packaging with guile yet? *wingo doesn't know :) ijp was mostly the one working on it and it's been a while <davexunit>it seems that we need some hackers that want to hack on guildhall <wingo>there isn't a gnu instance of it anywhere that i know of, though some people have managed to use it <wingo>but i got distracted by other things and didn't push to completion :/ <healer>how are guile packages beign distributed at the moment? <davexunit>in various other package managers for distros <davexunit>unfortunately, debian doesn't really have any guile library packages <healer>will guile be ready to replace emacs-lisp this year? <wingo>maybe! bipt is working on it again :) <mark_weaver>healer: there's no plan to replace emacs lisp. there's a hope to replace the *implementation* of emacs-lisp that in emacs. <healer>In the header for guile scripts, the \\ meta-switch expects all the following switches on a single line before the !# line, the script fails to run if arguments are on multiple lines, is this expected behaviour? <ArneBab_>wingo: that bipt is working on it again is great news! <wingo>i hear bipt is currently in progress of rebasing that massive patch set on top of current emacs git <paroneayea>as everyone already knows, I am super excited about guile-emacs ;D <ArneBab_>wingo: guildhall worked pretty well for me, but IIRC there were some rough edges (some command options I had wished for). I think it would be great to have it packaged. I cannot promise that I’ll get to look into it this week, though. <wingo>ArneBab_: guildhall needs a champion :) if you end up doing that you would be a hero to the masses! <ArneBab_>I’d love to, but I have to be realistic with my free time, so I’d rather not promise it ☺ (though it’s good to know that work on it would find support) <ArneBab_>but firstoff I have to move home… (family’s waiting) *paroneayea sends an email to guile-user about stdin, stdout, stderr <paroneayea>not sure if it should have been a guile-devel post but I don't think I know if there *is* a solution I'm missing :) ***michel_mno is now known as michel_mno_afk
<davexunit>since Guile is GNU's extension language, technically <ArneBab_>I guess for that, we’ll have to create the bindings. <ArneBab_>or rather: someone will have to add Guile for everything. And lilypond needs to get fixed, because that is currently the worst possible user story… <paroneayea>ArneBab_: how far off is lilypond from 2.X support? <ArneBab_>paroneayea: I could build the gentoo version with guile2, but it could tot compile a simple example sheet. <davexunit>IIRC, lilypond heavily relies on a quirk in guile 1.8 related to macros <davexunit>guile 2.0 fixed the issue, and lilypond breaks. <davexunit>I think it had something to do with evaluation order. referencing macros before they were define. <paroneayea>soon that tracker will be garbage collected with the rest of google code :( <zacts>indeed, google code will be no more soon <davexunit>github does have the advantage of being profitable, at least. <paroneayea>(Which means they probably are growing a bunch in expenses with hope that profit will go up likewise?) <davexunit>"let us help you build an unsustainable business that Google or Facebook or Apple will buy" ***first-healer is now known as healer
<wingo>mario-goulart: are you on pre-4.4 android? <wingo>the site only does tlsv1.2 fwiw <wingo>dunno if that is the right thing but there are some UAs that don't understand tlsv1.2 <wingo>what UA? works for me with debian sid and epiphany <mario-goulart>wingo: works on iceweasel. Chromium shows "This webpage is not available" <wingo>could be that chromium there doesn't do tlsv1.2 <paroneayea>davexunit: I wonder how hard it would be to get guile bindings for telehash using the FFI <paroneayea>I guess the main thing would be needing a bridge for the json stuff. <paroneayea>davexunit: a peer to peer system for communication, basically sending around json snippets. <paroneayea>main person is the main person behind xmpp/jabber <paroneayea>davexunit: yeah. I wonder if they have their own json representation inside telehash, and if it's so hard to serialize to a string to their structure, if they provide something <wingo>we need a fast bit-count primitive. <wingo>we need native compilation :P <wingo>i guess we should follow common lisp here <civodul>paroneayea: that looks pretty cool, indeed <paroneayea>note: git-timemachine doesn't like to run on my auto-commit-on-every-file-change orgmode git repo :) <paroneayea>given that I'm editing my main orgmode files all the time