<catern>has anyone backported a recent guile to debian wheezy? (shameful question, but I have these wheezy servers and I want to get guix on to them...) <mwette>I'm working nyacc documentation today. I think this will take a few weeks. <amz3`>catern: installation of guix binary is not linked to the host system, you can install it on any system i think <amz3`>I just sent an epic mail to the gnu system mailling list <amz3`>“Let's build a free self-sufficent and sustainable space-ship for a one way trip to another habitable planet with the goal to settling a colony.” <catern>amz3`: that's only if one uses the tarball approach to installation, which bundles a guix with it <catern>amz3`: but I want to actually make guix as a Debian package <catern>and that requires backporting a recent guile <nomad>is possible change the guile default prompt to other chatacter such as λ> <paroneayea>nomad: hm that's interesting... I don't think so, by my reading <paroneayea>modules/system/repl/common.scm's repl-prompt seems to have the format string hardcoded there <paroneayea>ACTION doesn't understand guile's internals much yet but sometimes knows where to look, on the scheme side... :) <nomad>paroneayea: jeje thanks for the tip :P <nalaginrut>janneke[cm]: I think it's good so far, just network is slow for me <nikita1>nalaginrut: what do you think: should the number of db connections be limited (as wqlen) and the next connection request should be pending for the next free connection slot, or the number of simultaneous connection may grow infinitely? <nikita1>in the first case the pending task should be somehow scheduled.. ***brendyn is now known as brendyyn
<amz3>it sound as if I was megalomaniac <brendyyn>Contrary to my expectations, filter-accumulate* runs nearly twice as fast as filter-accumulate. Is there any particular reason for that? <amz3>apparently my space travel project is off topic <ArneBab>amz3: your project is roughly a century of European research funding :) <ArneBab>I guess the way to start is to do one thing towards the goal — most of my experiments failed when I set perfect requirements first. <amz3>ecraven: what's the difference between both, I read about lojban, my understand is that it's less popular <ArneBab>ecraven: that’s actually a pretty hard question — there are several languages which might fit. <amz3>ecraven: any advantage to use lojban? <amz3>ecraven: indeed that's an issue imo <ArneBab>ACTION speaks some esperanto, it’s not mangled latin but rather a mix of European languages with a simple grammar. <ecraven>ArneBab: sorry, I was trivializing.. but lojban has very different goals (and roots) than esperanto <ecraven>"global" language than esperanto (which is very western-centric) <ArneBab>my core problem with lojban is that I don’t know whether it will actually work out: Are there tests on learnability by children). <ArneBab>(on the other hand, lojban has been on my to-learn list for over a decade) <ecraven>I think there are a few native speakers. but if you have fluent speakers, then it is learnable (and there certainly are some!) <brendyyn>Robin Powell spoke it to kids kids. Not sure if he still does. <ArneBab>ecraven: do you have nice spoken examples? <ArneBab>lojban — I could actually speak some toki pona at some point <brendyyn>I think once a week there is a mumble chat for lojban practice, or other voip software <ecraven>#lojban would know about good examples <amz3>my understanding is that the GNU project will end up doing similar the project I posted, they don't know yet that is all <amz3>what's the point of building software to sell vegetables? <amz3>I mean not all software aim a selling vegetables but... <amz3>to get started I need to build a software that works and has success, then maybe people will start listening <amz3>at least now I have a goal <amz3>someone on #gnu asked if I was Elon Musk x) <amz3>brendyyn: my post is not a joke <brendyyn>I don't wish for Esperanto to be any kind of official language, since I don't think it was ever well designed. <amz3>I am not sure about esperanto, what I want is language that is international <ijp>brendyyn: I think the only official use it has is a university in san marino <brendyyn>What I want is real time translation that doesn't suck. Unification comes with a dark side. it's one way languages get wiped off the map. <ArneBab>amz3: easy to learn it is, and it works for many people. I actually started learning it when playing a science fiction play-by-email <ArneBab>but I guess that’s somewhat off-topic here. <paroneayea>rekado: btw, I did snarf your multipart stuff into guile-webutils <paroneayea>rekado: I think I'm going to make it so that instead of bytevectors for the part-body it's ports, which for now are just ports wrapping bytevectors <paroneayea>but eventually it'll be able to also write things out to files if it turns out that it's reading a lot of information <paroneayea>eventually I'd like to allow the parsing procedure to also accept ports, alongside bytevectoras as it currently does <rekado>paroneayea: I’m glad it was useful enough to be snarfed :) Using ports sounds like the right idea. <rekado>when you make a release I’ll be sure to use it in rcas-web. <wingo>paroneayea: it throws an error at runtime <wingo>some file descriptors are seekable, some are not :) <paroneayea>so I guess if I'm parsing the multipart message off a port that's not buffered and/or is non-seekable for whatever reason <paroneayea>there's the part where you read-ahead to see if there's a match and I guess <wingo>the port is buffered byt not seekable <wingo>you can always unread-char or whatever <wingo>or rather, the port is not buffered by default <wingo>but you can make it buffered <wingo>but lots of unread-char leads to inefficient parsers; i wouldn't recommend it as a fundamental thing <paroneayea>okay I guess that doesn't seem like the right approach :) <paroneayea>I guess that does mean doing some buffering myself then? <stis>wingo: got a question: should we commit the simple case of -g0 first and then make a guild strip funciton <stis>sneek later tell wingo: got a question: should we commit the simple case of -g0 first and then make a guild strip funciton <paroneayea>that moment when you realize your web application is stupidly vulnerable to the "../" hack attack <amz3>oops! I thought about that but don't remember if i checked, i will verify it right now <daviid>paroneayea: doesn't guix have a bunch of path related procedures that are as good as python? actually I did think to snarf most of these so they become part of guile-lib :) <paroneayea>daviid: I don't think you can just snarf them into guile-lib though because guile-lib is LGPLv3 :) <paroneayea>you'd have to ask the copyright holders about relicensing them since Guix is GPLv3 <nalaginrut>wingo: oops, guile3 now? I guess it's time for AOT, right? <daviid>ah, licenses, guile-lib has the 2 licenses, and specify that anyway it is a 'per module' matter, so that would be fine I think <daviid>paroneayea: anywau, when you think guile misses something comared to python, let's talk about it and if apropriate we write a module in guile-lib ... <paroneayea>daviid: I see, I guess guile-lib is more GPLv3 than it is LGPLv3, according to guix at least :) <paroneayea>daviid: yeah I might have something to contribute to guile-lib <paroneayea>iirc there's *some* OS out there that uses multiple characters for a path separator, though I forget which it is. is it OS/2? <paroneayea>well I'm not gonna support OS/2 so it doesn't matter I guess :) <daviid>:) I only use GNU linux so I can't answer that quiz <daviid>stis: nice to see one of finally grabbed a task from wingo's list, this is awsome! <stis>nothing is stoping anyone more to step forward. It would be great to socialize around a programmoing task. <daviid>davexunit: ah yes I remember, but some of your code got ito guile proper no? <stis>not hitting wals is booring! <nalaginrut>wingo: I've read the dissertation of puzzle solve for register allocation. I think it's based on SSA, we have to make it for CPS first. <stis>but you should manage to pulverize some of them so there shouljud be a balance. <stis>ACTION is chewing on elf format issues, .... still a great wall of oignorance remains <stis>-g0 should be done, but reading in an elf strip som segments and write it again needs understanding. <stis>actually I think that one strip sections and not segments. <pookleblinky>Do you think guix is gonna finally be guile's killer app? <jlicht>pookleblinky: it isn't already? ;-) <pookleblinky>jlicht: it's already by far the most interesting use of guile