<sirgazil>amz3: yes, I'm interested in scheme in the browser, so thanks for the link. <wingo>but it's a really good explanation of the state of the art <wingo>still if guile had something like reagents and if fibers were mature, that would go a long way <wingo>futures are pretty great for dataflow parallism like in the par-eval sense <wingo>but i dunno if we need asyncs or something too; that would have to be based on fibers or something like that <wingo>i don't know how much of a story guile can have without a native concept of lightweight threads <wingo>i also don't know what the interaction is/should be between fibers, interrupts, and mutexes <joolean>wingo: Did you happen to see my latest ECMAscript patch? <paroneayea>oh, patch! also cool, though I thought you said branch :) <paroneayea>joolean: ecmascript support can use more love though :) <joolean>paroneayea: Well, my dream is to be able to start rolling portions of this: https://github.com/tc39/test262 into Guile’s ECMAscript test suite. So I’ve got a series of patches to get Guile’s ECMAscript foundation to a point where it can bootstrap that test suite. <wingo>joolean: i haven't seen it but tx for the reminder :) <wingo>tfw you look at the docs and they are terrible <joolean>surely you don’t mean the Guile docs <paroneayea>guile docs are great outside of the tutorial section :) <wingo>looking at the guile docs yeah <wingo>also how character sets are somehow deemed a simple data type <wingo>and of course the worst, "scheduling" <wingo>where somehow parameters have to do with scheduling (???) ***karswell` is now known as karswell
<daviid>weird: M-x run-guile -> GNU Guile 2.1.4.87-fcb43, (use-modules (cv)), all fine, then open a buffer with some test code, go to the end of a prodecudre def, C-x C-e -> (scm-error misc-error #f "~A ~S" ("no such language" objcode) #f) <daviid>this works fine on my machine, and fails on another, I wonder why <daviid>have to go, bbl. if anyone has a clue, please ping me, I'm running the latest geiser and same guile version on both machine ... <wingo>really wish we could junk dynamic states <janneke>wingo: there's something in the commit log <dsmith-work>I bet that dynamic-root stuff has been in there since SCM. <wingo>janneke: it was more of an exclamation than a question <wingo>omg you can pass a dynamic state as the second argument to eval <wingo>i wonder how parameters should work with fibers... <wingo>of course parameterizations local to a fiber work as you would expect <wingo>but what should the initial parameterization of a fiber be? <wingo>should it inherit from the fiber that made it or should all be the same? humm <paroneayea>hm right, there's no way to read from stderr in open-pipe*, right? <paroneayea>I think I've asked this before and the answer is still "nope" <paroneayea>so much for losing patience and writing this backup script in guile... not sure how I'd shunt rsync errors! <paroneayea>I mean I could use librsync but I don't have a lot of time ;P <wingo>simplest is to run bash in the (ice-9 popen) and pipe the stderr somewhere :/ <paroneayea>ACTION puts "add better popen to guile" on never-ending todo list... <wingo>paroneayea: did you see asyncs.test in guile? pretty neat <wingo>you can preempt and resume now <wingo>anyway, thought you would be interested <paroneayea>wingo: I'll be almost certainly writing stuff using these tools shortly <paroneayea>wingo: I need to write a activitypub test suite that tests other peoples' implementations <paroneayea>and since I've already written pubstrate in guile...