IRC channel logs
2017-07-27.log
back to list of logs
<paroneayea>either I need to figure out how to proxy each read/write to the underlying port and watch carefully for -1 <paroneayea>or I need a way to not "call through C" to the supplied read/write function, which means setting scm_read and scm_write, but apparently that's currently done on the port type rather than the port itself <happy_gnu>I am trying to pass as arguments "and" or "or" <happy_gnu>but when i call the function with them i get a syntax error <happy_gnu>"unknown location: source expression failed to match any pattern in form and" <daviid>happy_gnu: you can't do that because and and or are macros, not functions <daviid>what you can do is pass a symbol and use case in the body <daviid>(define (comp op foo bar) (case op ((or) (or foo bar)) ((and) (and foo bar)))) <daviid>not sure what yu are trying to do though <happy_gnu>i never try to pass them as arguments, couldn't understand why this didn't work <happy_gnu>there are 2 functions intersects? and subset? they are really similar just change the "or" and the "and" <happy_gnu>so I thought of making one that can cover both <happy_gnu>this is not part of the little schemer but I am trying to do everything that comes to my mind <daviid>happy_gnu: ok, good to experiment yes <daviid>i wonder why reduce args were defined in that order, I'd rather reduce something using a procedure and an initial value ..., just curius <happy_gnu>daviid: ohh they do not use them the way I did :/ <happy_gnu>I just saw that 2 very similar functions were only different for and "and" and a "or" <happy_gnu>so i tried defining one and just define them using the one that groups them but I didn't know I couldn't pass "and" and "or" as arguments :V <janneke>paroneayea: ouch, sorry to read you're stuck...good luck with that! <magnicida>is there a way to produce/compare SCM_UNDEFINED from Scheme directly? <ArneBab>paroneayea: you could stray from the cooperative model and spawn a receiver pthread just before you send the request <paroneayea>ArneBab: I could, though I'd need a good way for it to send a message back to the actor <paroneayea>I could do that more feasibly if I finished switching 8sync on top of fibers <paroneayea>maybe that is the fastest way to make this happen <paroneayea>but it's a pivot in strategy, and it's still leaving a problem open <ArneBab_>paroneayea: doesn’t ActivityPub send a message to the test when it gets a reply from the other server? <ArneBab_>(to make this even cleaner, you could start a second process which only listens for the request from ActivityPub and provides the expected result) <paroneayea>ArneBab_: yeah, so here's the situation that's happening <paroneayea>but we're also logged into the remote server, using the client to server API <paroneayea>where we can post messages logged in as that user <paroneayea>and submit that to our outbox endpoint of the user we're logged in as <paroneayea>when we do the http request, in a cooperative model where https blocks, we're waiting for the response <paroneayea>the server getting it, we have no control over how it's written <paroneayea>and it may actually try federating the message *before* we get an HTTP response <paroneayea>but the server is blocked on the initial request <paroneayea>anyway, yes, if this was running in a separate process or thread, it wouldn't be a problem <paroneayea><paroneayea> ArneBab: I could, though I'd need a good way for it to send a message back to the actor <paroneayea>I wasn't referring to messages as in terms of AP messages <paroneayea>so, layering 8sync on top of fibers would fix this, because each actor could run in its own thread, and yeah you could have a worker process to do requests <happy_gnu>paroneayea: I've learned a lot of lisp since the last time we talked and I was trying to look for your game to see if I understand it better now <catonano>happy_gnu: the video is on the 8sync web site