IRC channel logs
2014-03-06.log
back to list of logs
<artyom-poptsov>civodul: I'm afraid I haven't figured out why your patch leads to the bug in sssh yet. But I think it's matter of time. <civodul>yeah, would be nice to figure out ;-) <civodul>i have code that pipes to lsh processes, and i'm more comfortable with using Guile-SSH instead :-) <artyom-poptsov>As far as I understand, you're going to use it for distributed build of packages. <artyom-poptsov>I saw a recording of your talk on GNU Guix, but as far as I remember there was not much information about this topic. <artyom-poptsov>There is information about ongoing work on integration of Gnunet with GNU Guix (or vice versa) to share built packages between peers. So it seems that Gnunet and SSH will be used for different tasks. <civodul>artyom-poptsov: yes, that's a different thing <civodul>SSH is used to offload builds from one machine to another <civodul>typically on the build farm: there are several machines, and the "master" decides what to build, and possibly offloads stuff using SSH <artyom-poptsov>civodul: Hmm, it seems that distcc does similar job -- it shares compilation tasks between specified hosts, except that it does not provide authentication mechanisms (AFAIK). <civodul>artyom-poptsov: well here the thing is integrated with Guix, and Guix controls the build env., knows what's missing on the target machine, etc. <civodul>so conceptually it's similar, but in practice it's very different ;-) *civodul didn't know about SRFI-71 <civodul>i don't see the improvement over SRFI-11, though <b4283>indeed 75 and 11 are very much alike, wonder what's the difference <ArneBab>nalaginrut: I just googled for “let values” to find an SRFI I can use <ArneBab>nalaginrut: but I now added support for srfi-105 in wisp <nalaginrut>ArneBab: well, I mean sweet-expression, I didn't mean let-values ;-) <ArneBab>nalaginrut: because sweet expressions became much too complex for my taste <ArneBab>nalaginrut: I contributed to them (on the mailing list), but when they grew hairs like using $ and <* *>, I realized that they had lost the beauty which fascinates me about lisp and scheme <nalaginrut>ArneBab: hey~out of topic, how do you generate that slide on the web? <ArneBab>they are special syntax to simplify some cases - which in my opinion only arise because they cannot continue the argument list after calling a function and because they try to avoid all parens instead of just removing the superfluous ones *nalaginrut don't want to waste too much time on syntax expression, but programming mind <ArneBab>I don’t completely understand that sentence <nalaginrut>I mean I have less interest on these expressions style <nalaginrut>sweet expression or wisp, anyway it's personal taste <nalaginrut>I'd like to think more about the programming mind <ArneBab>I am interested in them, because I remember how much easier it was for me to learn python than to learn C, and that scheme looked completely alien to me in the beginning, which actually hindered me when thinking about code <nalaginrut>yes, I don't think it's useless, just not my cup of tea ;-P <nalaginrut>I have a party tonight, see you guys tomorrow ;-D <ArneBab>civodul, b4283: I think SRFI-71 saves one level of parens compared to SRFI-11: let ((a b c (values 1 2 3))) vs. let (((a b c)(values 1 2 3))) <ArneBab>civodul, b4283: for the wisp-example this was an advantage. *wingo chatting with the fine chicken folk on #chicken <davexunit>be sure to crack a few yolks while you're there. <antoineB>do you know the repo for guile emacs? (asked the same question on #emacs) <antoineB>do you have experienced racket? (i only use racket as scheme) <mark_weaver>I know about racket, of course, but I've not used it.