<jcca>ft: Hi Can I do (define serial (open-output-file "/dev/ttyACM0")) <jcca>ft: (configure-termios...) <jcca>ft: (display "message" serial) <davexunit>guile-2d has officially been renamed to "Sly" <ft>jcca: You might need to be able to read from the file as well to be able to use ‘tc-get-attr!’ - I don't know because I didn't try with output-only files yet. <ft>jcca: Can you elaborate? :) <ArneBab>davexunit: are you sure about that name? <ijp>guile has the advantage of not being in your typical persons vocabulary <ijp>as an aside, if I ever wrote a scheme compiler, my plan was to call it baldrick <ijp>(an *actual* compiler, not the toy ones I do anyway) <ArneBab>davexunit: guile actually is in many peoples vocabulary: Guile Theme goes with Everything <ijp>Guile there is a proper noun <ArneBab>but the meaning is a very different one - and one nown internationally ☺ <ArneBab>ijp: alternatives with a similar but positive meaning are wily and astute <ijp>wily has unfortunate coyote connotations <ijp>anyway, naming this thing isn't my decision *ijp sticks on SF2 theme songs <davexunit>didn't realize my choice of name would spark a conversation like this. <davexunit>I think sly is okay. Some like it, others don't. Can't please everyone but it's good enough for me. <ijp>guile 3d would have worked until you added 4d support <ijp>guile superhappyfuntime ***siel_ is now known as siel
***karswell` is now known as karswell
***_zxq9_ is now known as zxq9
<janneke>do we have a procedure/module that resembles /usr/bin/diff ? (esp diff -uw) ***endou__ is now known as endou_
<wingo>alexei___: tx; can you drop a brief mail to bug-guile@gnu.org? my tree is borked atm, and that will make sure we don't forget <wingo>no subscription needed; a mail to bug-guile makes a bug report ***_zxq9_ is now known as zxq9
<ArneBab>davexunit: But in the end it’s your tool - and there’s a limit to the effect of names (if sly works really well for many people, the name won’t really hurt that much). <alezost>if it wasn't too late, I would suggest "gruile" ("gr" for "graphics"), besides it sounds a bit like "Grail" *janneke just hacked a (define* (diff a b :optional (options "-u") (virtual-name-a "a") (virtual-name-b "b")) <janneke>and wonders if and why he duplicated some efforts, or else where he could dump it ***nalaginrut_ is now known as nalaginrut
<wingo>(except they are sparse, unlike clojure's vectors) <alexei___>((< max-fixnum val) max-fixnum) in (define-syntax-rule (clamp-range val) is probably a typo *davexunit is trying to port chicken's lru-cache <davexunit>want to cache game assets loaded from disk and regular memoization will just keep eating memory. <wingo>you don't want to use mmap or something similar? <wingo>with MADV_DONTNEED to free and MADV_WILLNEED to page back in *wingo very pleased with recent flow analysis stuff; needs a writeup to <wingo>happy to have landed most of it today <davexunit>I just want a cache with some upper bound because the cache will exist for the entire duration of the game. <janneke>is there a neat way to avoid writing this lambda; i.e., discard : (lamba (x) (some-fixed-procedure)) *janneke looks for some kind of compose that junks a parameter, or something <janneke>i have a list of `((key . ,proc) (key-2 . ,proc2)) pairs <janneke>the procedures are invoked with one parameter <janneke>sometimes, the "proc" does not need the parameter <janneke>i now changed the library code to test for procedure? <janneke>and otherwise use the fixed value directly <davexunit>I think this has come up before: there is no hash-table-size procedure for guile's hash tables. <ijp>I forget the reason why, but you can compute it fairly simply <ijp>e.g. (hash-count (const #t) h) <daviid>davexunit: hello! just reading the log and searched for sly, found a s/w already named sly [didn't check if it's alive but thought you would like to know]: "Introduction. Sly is a JavaScript library for advanced one-directional scrolling..." http://darsa.in/sly/ <davexunit>it's not a major clash. different library, different language. <daviid>davexunit: i think install.org file is outdated, it still refer to figl in dependencies <daviid>the "Freeimage is to used to load..." is incorrect, I think <davexunit>glad to have someone reading what I've written <daviid>I would remove debian subtitles and add a sentence in the beginning saying that the install command propose are based on a debian ... it really is redondant i think <daviid>s/command propose/commands proposed <davexunit>daviid: are you looking at the latest code in master? my INSTALL.org has guile-opengl now. <daviid>i picked up the first guile-2d link googleing <davexunit>I modified the README on github so that there's bold text that directs you to gitorious. <daviid>you should update [maybe] the github page to point to the new one... <davexunit>I only use github to work with upstream projects that use github now. <davexunit>my own projects are on gitorious because it's free software. <daviid>so, i can not install sly [for the same reason we talked about..], just reminding <davexunit>ijp: thanks for the hash-table-size suggestion <davexunit>daviid: one sec, I can attempt to help again <daviid>davexunit: i'm curious why users hit this problem [often] and also would to use the autotool 'schema' you're using for kisê :) <daviid>./configure: line 2384: GUILE_PROGS: command not found <daviid>anyone can help with this? [guile-sdl install]: ./autogen.sh --prefix=/opt -> ./autogen.sh: 25: ./autogen.sh: snuggle: not found <davexunit>daviid: what is the output of: find /usr -name guile.m4 <davexunit>ahhh well that explains why I've never encountered this before. <daviid>remember that we've been through this with civodul already, guile,m4 is installed <daviid>in my case it's here /opt/share/aclocal/guile.m4 <davexunit>yeah I know, but I don't remember all of the details <daviid>my ACLOCAL_FLAGS is properly set as well: cho $ACLOCAL_FLAGS -> -I /opt/share/aclocal <daviid>see this guy is terrible [the sdl guy] because he pretends you can use autoconf, but makes it [hiddenly] depending on a s/w he wrote <davexunit>I tried packaging guile-sdl for guix and gave up <daviid>SNUGGLE: Neutralize Uncharacteristically Grotesque GUILE Library Entropy! <davexunit>daviid: man, I don't know what's up with this guile.m4 thing. <davexunit>the mailing list might be helpful, mark weaver could see it and might know what the issue is. <daviid>snuggle snuggles me, i don't even want to install that crap <davexunit>oh wow, that was back during the GNU30 hackathon <daviid>yeah [wanted to have lunch with you and mark, but then i live about 10.000km :)] <davexunit>the (system repl coop) module was conceived that day. <daviid>by the way did you look at the clutter api? <daviid>i wish mark would write me: david anything i can do to help you with guile-clutter? :) [oh maybe 1 day he will :)] <daviid>but guile-clutter needs help since ... 2012 :) <daviid>gdk_pixbuf_get_pixels returns a guchar *, clutter_image_set_data wants a const guint8 * <daviid>so, I (wrap-opaque-pointer! ws "guchar*") and commented the ignore gdk_pixbuf_get_pixels so that function is now wrapped as well, and calling returns [for example]: (get-pixels pb) -> #<gw:wcp <guchar*> 0x3ccf930>, which fine *daviid is speaking loudly, explaining to someone else, sometimes, unexpectidely triggers other route(s) for a solution... <daviid>now, we don't wrap cogl [yet :)], and clutter_image_set_data wants a CoglPixelFormat, I wonder what would be best here <daviid>should I add a colg types entry in Makefile.am ? or create a nasty hack just for that cogl type? <daviid>or add a cogl directory to the clutter branch [which would trigger a lot more work] *janneke is happy: found const ***ijp` is now known as ijp