IRC channel logs
2017-03-29.log
back to list of logs
<wingo>has anyone implemented zlib in guile yet <wingo>or DEFLATE compression/decompression rather <Apteryx>How do I split a string? I want the tail of a path which comes after ".../man/" <civodul>wingo: i think Industria has a pure Scheme implementation of it <civodul>otherwise there are bindings to the C lib <Apteryx>string-contains looks like a contender <wingo>ah right. i should take another look at industria <wingo>i want to see how fast/slow guile is <daviid>lloda: heya! do you have a fast m,atrix multiplicatin algo somwehere? Coppersmith-Winograd maybe? <lloda>daviid: I use cblas, my own bindings <lloda>I have published ffi bindings, unfinished, but the xgemm functions are there <lloda>I always thought winograd curiosity, since performance always seems to be dominated by cache issues, ime <lloda>so much work goes into tuning these routines, it's pointless to write your own <daviid>lloda: thanks. guile-cv which binds vigra, have images reprented by f32vector, would that work? <lloda>that would be sgemm/sgemv, yes <lloda>cblas doesn't support all possible strides, so you have to be aware of that <lloda>my ffi bindings don't support negative strides, either <lloda>cblas does those in a strange way :-p <lloda>iiuc there's only one #f value, so yes <Apteryx>How do I know what kind of exception I am dealing with? The error is: "ERROR: In procedure symlink: File exists", and I'd like to catch and handle this exception. <Apteryx>Is it a 'system-error? How am I supposed to be sure? <Apteryx>wingo: Thanks for the above answer! I had to use string-contains since I'm interested in spliting based on a substring rather than a character. <janneke>ACTION is voting for a GNU standard time <davexunit>GNU standard time is US eastern time since that's where the FSF is <Apteryx>What do I put in my exception handler to do... nothing? <Muto>Maybe I'll write a virtual pet in Guile. <wingo>aaah, i think i can include lots of nice things in a guile sandbox: datum->syntax, eval, and so on <Muto>OrangeShark: That would be amazing! <void_pointer>Wonder if the new syntax stuff will make debugging syntax-case macros easier <void_pointer>Had just made my first syntax-case macro this last weekend and it was quite a bit of work to get it working and debug it <void_pointer>And worst of all, once I got it working, turns out that it was insufficient for the job and had to come up with another solution (basically, another syntax-case was acting before it) <void_pointer>Will be interesting to see the new syntax objects when they are done <lloda`>void_pointer: I just suffered through some syntax-case pain recently. <lloda`>iiuc the new syntax objects are just meant to make them a distinct type instead of vectors or whatever the were before. No new features (wingo correct me) <lloda`>I think just copying a few of the racket facilities like syntax->e, syntax->list, (define-syntax (m stx) ...) etc would help, if only so that the racket tutorials are easier to follow <lloda`>b/c they're the best tutorials I've found, but they're aren't 100% standard rxrs <void_pointer>Haven't actually used the Racket tutorials that much because of them not being RNRS exactly <lloda`>a couple of things are different but they're very minor <void_pointer>Page looks familiar. I think I glanced at it half a year ago