IRC channel logs


back to list of logs

<cky>mark_weaver: Would you be interested in a shell account on a Linode instance so you can have an always-on IRC connection? Now, the instance doesn't have a lot of memory so I can't promise Emacs will rock on it (since I IRC from a weechat-over-tmux session instead), but at least you won't be at the mercy of your ISP. Granted, that means you'd have to run Emacs in -nw mode, which may be unacceptable to you.
<mark_weaver>cky: thanks for the offer, but I could do this on also. I might have to do that, but I think I'll first try routing through tor and see if that helps.
<mark_weaver>though my experience with connecting to freenode through tor is also spotty. some days it's simple not possible, because of a problem on the freenode end.
<sneek>Welcome back dsmith, you have 1 message.
<sneek>dsmith, mark_weaver says: sneek thinks it's early May.
<dsmith>sneek, seen dsmith
<sneek>dsmith was here Dec 09 at 12:18 am UTC, saying: sneek, seen dsmith.
<dsmith>sneek, later tell mark_weaver Ok be root, reboot all is well
<sneek>Will do.
***fangism is now known as fangism-hungry
<nalaginrut>morning guilers~
<davexunit>good morning nalaginrut
<davexunit>how goes documenting artanis?
<nalaginrut>davexunit: I've done part of the CSS, and I'm filling the content
<nalaginrut>it's not updated yet
<nalaginrut>davexunit: how about write this game?
<davexunit>hah, looks impossible
<ijp>nice, it's available for linux
<ijp>oh wait no, googles sidebar lied to me
<zacts>hi guile hackers
<mark_weaver>daviid: if the subprocess exits without outputting anything, then the 'read-line' call will return EOF and proceed. but you could test this yourself by just changing the command to "sleep 10" or some such.
<mark_weaver>daviid: based on the process listing you provided, it looks like the subprocesses are sleeping. I recommend attaching a debugger to one of those long-sleeping subprocesses to find out what it's doing.
<daviid>mark_weaver: tx, i did figure out in the mean time. it appears that these processes go to sleep because og imagej, not clojure. i need to further investigate but it is probably because of imagej 1.xx depends on an X server, somehow it does not understand/deels with "ssh -CX ..."
<mark_weaver>sounds plausible.
<daviid>and i mostly work remotely, too bad, but i think i can [and have to] find a solution
<daviid>so far guile is perfectly fine, which is a good news anyway [because i'm the one who recomanded it so i fell very concerned that anything opn the guile side is always good or quickly debuged
*mark_weaver goes afk for a bit
<daviid>ok tx again
<nalaginrut>ArneBab: I want to customize a new style link, do you have any experience?
<nalaginrut>for an instance, all the [[] [artanis]] has a special id/class for a color
<nalaginrut>for now, it's a common a tag, without any class/id
<ArneBab>nalaginrut: you mean you want all links with that class look different?
<ArneBab>nalaginrut: or do you mean you want all links to look different?
<nalaginrut>I mean one of the links named 'artanis' looks different
<ArneBab><style type="text/css">a .artanis {color: #123456}</style>
<nalaginrut>I know what you mean, but org-mode seems don't give me the chance to specify 'artanis' class
<ArneBab>that then colorizes links of the type <a class="artanis" href="...">linktext</a>
<nalaginrut>is there any way to specify class/id of a link in org-mode?
<ArneBab>ah, ok - I can’t find anything so I guess there you will have to use js…
<ArneBab>but it might be better to hack the org-export function.
<ArneBab>(describe-function 'org-html-link)
<ArneBab>nalaginrut: just add an elisp source block at the top which replaces the link function when the content is artanis.
<nalaginrut>ArneBab: is there any example for the elisp snippet?
<ArneBab>nalaginrut: the org-html-link function seems to use attr_html
<ArneBab>nalaginrut: before going there, please give me a minute to test a possibly better way.
<nalaginrut>alright, thanks ;-P
<ArneBab>For single links you can simply use
<ArneBab>#+attr_html: :class foo
<ArneBab> [[]]
<nalaginrut>does this mean all the a tags have class foo?
<nalaginrut>not for me, not all a tags are artanis link, some are not, I want to set class 'artanis' to those a tag who link to artanis site
<nalaginrut>s/not for me/now for me
<nalaginrut>alas, it's bad grammar to mix these terms in a sentence
<ArneBab>will take a moment
<ArneBab>nalaginrut: that attr_html onyl applies to the next link
<ArneBab>I’m checking for a way to do this for every link which is named artanis
<ArneBab>nalaginrut: I assume that the right way to do this™ would be to use (org-export-define-derived-backend)
<ArneBab>and in that change the mapping (link . org-html-link) to (link . org-html-link-artanis)
*nalaginrut is trying...
<nalaginrut>ArneBab: nice! +attr_html is enough I think, thank you very much!
<ArneBab>nalaginrut: glad to
<nalaginrut>BTW, is there any good proposal for color scheme of documentation?
<nalaginrut>color scheme is very important because it's color Scheme!
*ijp gives nalaginrut an F
***_zxq9_ is now known as ^_^zxq9
***^_^zxq9 is now known as _zxq9_
<civodul>Hello Guilers!
<sneek>Welcome back civodul, you have 1 message.
<sneek>civodul, nkar says: ocaml's testsuite passes in /tmp. any idea why it timeouts while building as usual? do I need to increase the time before the exception is raised or something?
<tadni`>Am I missing something obvious? I'm trying to add a docstring to guile, and I have (define (word-lookup word) "xxxxxxxxxxxxxxxxxxxxx") and the attempted doctring just colors as a string in Emacs.
<davexunit>tadni`: the issue there is that the body *only* has that string.
<davexunit>which means that it's not a docstring, but returning a string.
<wingo>"By convention, if a procedure contains more than one expression and the first expression is a string constant, that string is assumed to contain documentation for that procedure."
<tadni`>davexunit: Even adding a trivial (print "hello world") to the body, still gives it typical string coloring. :^P
<davexunit>well yes, it will have the typical string coloring.
<davexunit>it's a string
<tadni`>davexunit: Hm, is there a way to specify this in the emacs-mode maybe then? I'm used to having a different color from typical strings -- due to Emacs Lisp.
<tadni`>Like my strings in Emacs Lisp are pink, but my docstrings are a Darkish grey.
<davexunit>you'd have to adjust scheme-mode somehow.
<davexunit>do share if you figure it out.
<tadni`>davexunit: Yeah, I'll make a note to do so. I still have to study today into tomorrow ... my last final is tomorow about 24 hours from now, but after that, I'm dedicating some major times this break fudging with Guile. :^)
<tadni`>I'm just too excited to wait. :^)
<daviid>i'm interested to, i'm a pathetic emacs user :) and dream that someone shares his/her enhanced scheme-mode [dark-mode based, i use black background] for guile: lots of guile scpecific macros and procedures are not coloured and not properly indented
<tadni`>daviid: Yeah, there's quite a bit of love Scheme mode needs for Guile proper.
<daviid>l'd love to receive help, somehow i love emacs but i don't invest in these tings :)
<tadni`>That or if we can't get upstream, we should fork to get a Guile-mode or something and bundle that with Geiser.
<davexunit>daviid: fixing identation is trivial
<davexunit>you'll have to do that yourself anyhow for each macro you may write.
<tadni`>Really, coloring some default macros that aren't is probably trivial too. It's just getting these changes upstream that I see problematic.
<daviid>tx, what about advanced colouring mode, anyone?
<daviid>i think we [guilewrs] should create and post .emacs extract for guile
<tadni`>.emacs extract? Like a blurb of an init-file, specifically for guile?
<daviid>yeah, my english is bad today, but i mean something a user can grab, add to his/her .emacs and editing guile source code becomes a pandora trip
<tadni`>daviid: I mean, ideally, we would just ship an sane defaults in Geiser -- no?
<tadni`>Or something similar.
<daviid>geiser? i mean emacs, not launching geiser [yet] would already 'know' guile, all macros, specific colours for class of procedures [goops, os related ...] something advanced by someone who love these things and knows emacs ...
<tadni`>Really, providing a unified Guile development system/environment, might be a good idea -- assuming we can't get upstream Emacs approval in a few years. As in, ship guilemacs, geiser, guile-mode (scheme-mode but with guile specific coloring and indendation), etc, etc.
<tadni`>Hopefully, we'd be able to get most if not all these things upstream though.
<daviid>davexunit: tx a lot, will look at it right now busy on something else, but as i dream of an advanced guile emacs editing mode i jumped in
<tadni`>Time will tell.
<tadni`>daviid: I'm saying, assuming we couldn't get Guile-specific changes in Scheme-mode -- it might be a good idea to ship a fork of scheme-mode, geared at guile, with Geiser. Since geiser is pretty much the goto development package for Guile.
<daviid>tadni, i'm not in favor of 'unique thinking' if that is what you are trying to say. i think that if an advanced mode exists, people, users, students that don't know scheme yet... will be pleased to use it and of course can change colour scheme following their wishes...
<daviid>i really mean guile specific, since guile is now almost an operating system , it has more procedures then you you can dream of, far behond other scheme... hey hey
<tadni`>daviid: Eh, I rather have a unique mode for Guile that defaults very nicely and not have to tweak to get everything working ideally -- and tell someone to grab that, than have them deal with scheme-mode (assuming we couldn't get such changes upstream) and leave not as good as an impression that it could with a new user.
<daviid>a customized emacs config for identation and colour, for beginners, would be nice
<tadni`>Seeing that Guile, is making pretty much gearing towards an OS-like environment having a go-to package is more ideal anyways. Most people seem to use Geiser for Guile development anyways and throwing in a mode that makes such a thing even more ideal... I don't seeing being an issue.
<tadni`>Again, we should try to get such changes via upstream scheme-mode though.
<tadni`>But if we can't there are existing avenues, to get such a thing to people, interested in Guile, in mass.
<daviid>i use geiser too, of course, fine if geiser brings, once loaded, an even better editing guile's source code mode, fine, but a good offer for a series of .emacs entries would be nice. anyway, said everything i wanted, have to work now
<tadni`>daviid: Peace.
<tadni`>But yeah, the short term goal would certainly to have something like that.
<taylanub>tadni`: (define (foo) "docstring" #f). otherwise, "docstring" is *not* a docstring; it's the body of the procedure, meaning the procedure constantly returns that string.
<tadni>taylanub: I have (define (word-list word) "bazbarfoo" (print "Hello World")) currently.
<tadni>I'm not going to worry about it too much though, because I should be studying today. :^P
<tadni>But yeah, that makes sense.
<amirouche>héllo guilers
<tadni>amirouche: coi
<tadni>amirouche: Greetings, in Lojban.
<tadni>.i coi la .amirouche.
<tadni>Okay, going to be mildly responsible and study a bit for my last final.
<tadni>Peace for now. o/
<dsmith-work>Morning Greetings, Guilers
<sneek>Welcome back dsmith-work, you have 1 message.
<sneek>dsmith-work, mark_weaver says: sneek thinks it's early May.
<dsmith-work>sneek: seen wingo?
<sneek>I last saw wingo on Dec 09 at 02:14 pm UTC, saying: "By convention, if a procedure contains more than one expression and the first expression is a string constant, that string is assumed to contain documentation for that procedure.".
<dsmith-work>sneek: botsnack
<sneek> 15:56:26 up 15:40, 0 users, load average: 0.14, 0.05, 0.06
<amirouche>tadni: I hope for you that it's a Lojban exam ;)
<tadni`>amirouche: It's an easy math class, but I haven't studied for weeks... I just caught up on about 6-7 weeks of homework two days ago. :^I
<amirouche>yeah, I know that very well... worst exams ever were distributed computing & AI... I regret it now, If I were more enlightened at that time I would have work more and do a PhD
<amirouche>I miss university...
<tadni`>I'm stuck at a community college, till next fall, so I'm probably not taking everything as serious as I should ... I still have like a 3.3, but still. :^P
<amirouche>you should... even if you are unlucky to have to work to get money, you won't have as much time as right now to actually learn things. Even by just following the curriculum, every time I check the curriculum of my university I get depressed
<amirouche>even if basic CS curriculum is very minimal, it is still better than when I was there (5years ago).
<tadni`>amirouche: I know, I'm pretty much at the breaking point ... I'm pretty sure that next semester is going to be the start of me seriously studying.
<tadni`>I got all my pre-reqs out of the way, so I can focus on the stuff I actually care about which helps. But anyways, this is offtopic. :^)
<stis>hwj guilers!
<stis>urgh, hej på er guilers!
<zacts>hi guile hackers
<ArneBab>tadni: why not bundle a guile mode with guile?
<dsmith-work>ArneBab: geiser is for racket too.
<dsmith-work>And others?
<tadni>ArneBab: Yeah, I suppose.
<tadni>Do people use it for Racket often?
<dsmith-work>jao: Do they?
<tadni>It seems like racket-mode might end up being extended to match geiser?
<tadni>Approximately, that is.
<jao>dsmith-work, there are racket users, yes. there's also support for scsh and chicken in the works.
<tadni>Yeah, thinking about -- guile-mode, if such a thing does manifest. It does make sense to ship it with Guile proper.
<jao>tadni, racket-mode follows the drracket philosophy or restarting everytime you modify something
<tadni>jao: Ah, so there's no active repl server or anything?
<jao>tadni, so in that regard is going to always be different from geiser...
<jao>tadni, yes, there's a repl
<tadni>Is this an explicit choice, or just poor design?
<jao>but you cannot re-define things piecewise
<jao>it's a very explicit choice
<tadni>Ah. That's a bit of a bummer.
<jao>racket people like it that way. they like to say "the top level is helpless" :)
<tadni>I mean, I suppose we don't want Geiser to victimize other dialects and/or variants of Scheme if we can help it. :^P
<jao>(but i'm sure racket-mode will be quite good: greg is pretty good)
<tadni>When I took my brief jump in learning Racket over Elisp as my first "real" Lisp, it was pretty nice to use.
<tadni>Definitely preferred it over Dr.Racket.
<tadni>But yeah, Greg makes some neat stuff. I'm using his static blog generator for my blog.
<jao>so am i
<tadni>How odd...
<tadni>Guile's Repl via Geiser has docstring coloring.
<tadni>Scheme-mode doesn't though. :^P
<tadni>Hm, maybe something wonky was going on with Guix/GNU Distro...
<tadni>Docstrings highlight on this box, in-general.
<rekado>is it possible to "open" a module, such that all definitions are splilled to the top-level?
<rekado>I'm again trying to test internal stuff with ggspec (because it's so much nicer than srfi 64), but I cannot access any values that are not exported.
<rekado>It only works if I comment the define-module expression (but I don't want to do that)
<rekado>(Is anyone here using srfi 64 for testing? I don't really like it at all, but maybe I just haven't seen the light yet.)
<DeeEff>I don't really use srfi-64 for testing, I wrote my own property-based test module, but I hardly write anything large enough in guile to justify using it.
***cluck` is now known as cluck
<ArneBab>rekado: you can switch to a module.
***fangism-hungry is now known as fangism
<ArneBab>(set-current-module 'name) I think
<ijp>alternatively you can access private bindings with @@ but for big uses that is inconvenient
<ijp>you can write functions/macros to do the full import, but it's not obviously better than switching here
<ArneBab>ijp: how do I get a module suitable for using via (set-current-module)?
<mark_weaver>(set-current-module (resolve-module '(ice-9 blah))
<mark_weaver>rekado: we use srfi-64 in the guix test suite
<mark_weaver>but the other thing about 'set-current-module' is that it has to be done at compile time, not evaluation time, so I guess you need to wrap it with 'eval-when' also. really an ugly hack.
<mark_weaver>I really think the better option is to make a separate test module, and have it import the module that's being tested.
<davexunit>rekado: I don't think it's a very good unit testing practice to refer to private variables in your tests.
<mark_weaver>sorry, I didn't read the backlog carefully enough. I see this is all about private bindings.
<rekado>davexunit: well, I want to test how public functions alter private state. The state is not exported, but the functions are.
<davexunit>I'm suspicious. do you really need that global state (assuming it is global)? is there a public procedure that may retrieve that state?
<rekado>This is for XMPP stuff and I need an internal cache of expected responses.
<rekado>for each expected response a handler can be registered that is to be called with the response.
<rekado>The API abstracts this all away.
<davexunit>couldn't you then exercise that code by simulating a response and making sure the right thing happens?
<civodul>it's better to avoid it, but sometimes "white box" testing is needed
<rekado>I could and I will, but the syntactic sugar is in a different module and I wanted to test this separately.
<rekado>there are two APIs, really. One really ugly that has callback hell written all over it and one with a nice monadic syntax (which desugars to the ugly one).
<davexunit>you can access a private variable like so: (@@ (module name) variable-name)
<rekado>davexunit: thanks, I'll try that.
<davexunit>an xmpp library for guile is a good idea.
<rekado>for the record: I was not successful with setting the current module.
<davexunit>I worked with the xmpp protocol a lot at my previous job
<davexunit>writing special purpose javascript clients.
<rekado>(not successful = the private variable was not available within the ggspec test suite)
<rekado>my goal was actually to write a pubsub service, but then I saw that there was no xmpp lib.
*rekado has to go now
<tadni`>davexunit is or at least was for a time, working on a guile-tox library/bindings.
*stis bites the dust and implements the iso-prolog module bastard solution, sooo ugly and incompatiable with our guile module system.
<stis>I need to maintain a separable name module maping living outside the normal module, (frons)
<tadni`>Is anyone using XMPP in mass, anymore?
<tadni`>I know and/or the fsf is hosting any instance of it.
<stis>anyhow I can now load quite may swi prolog modules, the compiler is slow though.
<stis>but running the code is acceptable
<stis>sorry for the spam, I just had to get it out :)
<tadni`>Hm, I wonder what would be the easiest way to play audio via guile. I guess I can just pipe a shell comand to mplayer or something.
<tadni`>Yeah, have no idea how to send a shell-command to guile. :^P
<tadni`>In Elisp, it's just (shell-command "whatever")
<tadni`>Maybe something like scsh would help in-route to something like this. :^P
<dsmith-work>tadni`: system, system*, and the pipe stuff
<dsmith-work>tadni`: All should be in the extremely fine manual
<tadni`>"extremely fine manual"?
<tadni`>Like a slim cut, of the manual?
<dsmith-work>Hah. No, as in awesome.
<tadni`>Ah, okay.
<tadni`>I thought there was like the opposite of the "developer's cut" for Guile, or something. Like "Guile's greatest hits." :^P
<dsmith-work>"fine" is one of the words sometimes associated with "RTFM".
<tadni`>dsmith-work: Ah, so it's midly sarcastic? :^)
<civodul>strictly speaking, the manual is not "fine":
<dsmith-work>tadni`: If you need the output of (or input to) something, see
<dsmith-work>civodul: THat's not the only thing that isn't fine.
*dsmith-work can't see his belt anymore