IRC channel logs
2016-02-16.log
back to list of logs
<davexunit>paroneayea: do you happen to have that haunt patch you mentioned a few days ago? <Jookia>So, calling a function like string-append that takes strings- I want to maybe pass in some arguments, maybe not. Is there a way to specify an expression that 'expands' to some arguments in a function that I'm calling? ie "(string-append "abc" ("def" "ghi") "jkl")" <sneek>Welcome back Jookia, you have 2 messages. <sneek>Jookia, mark_weaver says: you were right about the meaning of our license field. if upstream allows distribution under GPLv2 or later, then our license field should be gpl2+. lfam was mistaken about our policy. <sneek>Jookia, xd1le says: also idris has a primitive "Ptr" pointer type, so i was wrong again! don't know how i missed that. :/ <amz`>Jookia: maybe you are looking for 'apply' <amz`>(apply string-append (list "abc" "def" "ghi" "jkl")) <wleslie>someone is writing an Idris -> guile bridge?! <Jookia>amz`: Is there a better way to do that if I have multiple lists without being so verbose? <amz`>Jookia: you look for something like *args in python? <taylan>Jookia: (apply string-append '("foo" "bar" "baz" ...)) <taylan>Jookia: and (apply string-append "foo" "bar" '(...)) <taylan>Jookia: basically, last argument is a list that is "spliced" into the argument list to the procedure <taylan>oh, amz` already pointed that out, sorry -_- <Jookia>taylan: Oh I see, using it without (list could be good <taylan>Jookia: well I use ' aka quote there. that means the elements of the list won't be evaluated (e.g. as variable references), probably not really what you want <taylan>I just meant it as a quick example <taylan>can you give an example of what exactly you need? <taylan>you can also append lists with 'append' (apply string-append "foo" "bar" (append lst1 lst2)) <taylan>so if lst1 contains "baz" and "qux" and lst2 contains "fix" and "fux", the whole call becomes (string-append "foo" "bar" "baz" "qux" "fix" "fux") <Jookia>Is there a flag to make the Guile REPL support arrow keys/history? <taylan>Jookia: you can enable Readline in your .guilerc... <taylan>if you use Emacs you can also just use Geiser to interact with a Guile REPL <Jookia>That doesn't seem to work, pressing up gives me '^[[A' still <Jookia>Oh, it's ~/.guile not ~/.guilerc <Jookia>So now I have "(apply string-append (append '("test") '() '("ing")))", with the '() intending to be a function call in real code- Is there no way to have something cleaner? Is there perhaps an append that will flatten the '() or any variables inside it? <taylan>Jookia: you mean like (apply string-append (append '("test") (function-call) '("ing"))) ? <Jookia>Yeah, but without all the lists for things the '("test"). Perhaps it's unavoidable <taylan>you can put some leading arguments barely, but yeah, if you want to have some at the end you have to append them to some list. so e.g.: (apply string-append "test" (append (function-call) '("ing"))) <mark_weaver>Jookia: better to use (string-concatenate <list>) than (apply string-append <list>) <sneek>mark_weaver, you have 1 message. <sneek>mark_weaver, petter says: if 'cryptomount' in GRUB does nothing i think a good way forward is to insert the crypto modules with: "insmod cryptodisk" and "insmod luks". I'll work this into some documentation later. <stis>nalaginrut: yep, that's the standard hello and and Swedes are quite familiar with the IKEA history. <stis>Kind of cool that the company is build ouside the stock market. <davexunit>I'm going to fix the load path issue, apply your patch, and release 0.2 when I have time. <paroneayea>davexunit: sounds like a good potluck contribution! :) <davexunit>well, since I didn't bring anything else, why not? <amz3>the tests in `tests.scm` are not good, I sent it by mistake, you should check for unit tests inside the module culture.scm <amz3>has anybody GSoC ideas for this year? <amz3>it would be nice to complete the last GSoC.. <ArneBab>I have a crawler for the web of trust in Freenet, using sxml and match for the crawling. Do you think that would be a good match for the potluck? <ArneBab>give me an hour, then I’m at the right machine :) <OrangeShark>ACTION wonders if I should send one of my unfinished projects... <kristofer>is there a way to modify inputs with guix emacs interface? <davexunit>Mesa will be releasing in March with support for it. <davexunit>though I'm very concerned because apparently it's even harder to use than OpenGL. <davexunit>it's even more low-level, so it takes even more code to do simple things. <roelj>It should provide more "low level" access, right? <roelj>I wonder how that is going to look like <davexunit>I'm concerned that software freedom will continue to software in the graphics space due to this. <roelj>Yes, I'm going to have a look in the weekend. <davexunit>but I've only grown more skeptical and concerned about it. <roelj>About its impact on free software projects? <roelj>That it will take a lot of effort to gain anything from it? <roelj>Yes, I believe it takes a significant effort to get it right in the graphics drivers <roelj>And I really hope there are people who are willing to invest that time into it. <roelj>So that not only the proprietary drivers will use it <Jookia>Will Vulkan work on my GMA 4500 card? ;) <Jookia>My problem is that I probably won't be able to use any of this since my free machines are either using GMA 4500 graphics from 2008 or etnaviv graphics that support OpenGL ES 2. Maybe 3 <davexunit>but when the new Mesa comes out I will try it out on my laptop with intel GPU <Jookia>I hope they get some kind of software support, or maybe mapping to existing DRI <Jookia>I heard Vulkan was supposed to be simpler, like a RISC <davexunit>I worry about it being too difficult for someone like me to make a renderer with it. <Jookia>That's a valid concern. If you can't make something too proprietary for free drivers to implement, make it too complicated for free developers to use <roelj>davexunit: I'm not sure how much help I can be, but I do plan on diving into Vulkan over the next couple of weeks, so maybe I could help you with it. <roelj>But I have no experience with writing renderers. <davexunit>roelj: I would definitely like having some help making sense of Vulkan <davexunit>and seeing what we can do with Guile to bind to Mesa's implementation <roelj>davexunit: The first moment I can get enough time to really read about it is the 20th. Shall I contact you then, to see what we can do? <davexunit>roelj: I can't do anything particularly organized because I'm busy, too, but if we're both on IRC and hacking on it, let's chat. <roelj>davexunit: Great. Same for me.. <davexunit>one big change is that instead of GLSL, the shader language, they are introducing an intermediate language specification. <roelj>Alright. I really have to read up on things :) <davexunit>OpenGL is still going to be around, of course, but Vulkan has SPIR-V <roelj>But will it be efficient enough? (I don't know how much overhead lisp will bring) <roelj>Having a "higher-level" interface might make more sense.. <roelj>But that won't expose the greatness of course.. <davexunit>when I say "lisp" I guess I just mean "s-expressions". <davexunit>the result will surely look much different than any lisp we've seen <davexunit>potential step 1: directly represent Vulkar IR using sexps <davexunit>open question: would any existing part of guile's compiler tower be helpful for targetting SPIR-V? <ArneBab>sent my potluck dish! Happy Birthday Guile!