00:26 <tiehuis> scientes: not that i know of, i actually bought some cheap avr chips a few months ago to test but just have never got around to it

00:26 <tiehuis> also in regards to arguments, do you mean command-line arguments?

00:27 <tiehuis> in that case you can use argsAlloc, argsFree: https://github.com/ziglang/zig/blob/master/std/os/index.zig#L1925

00:27 <scientes> tiehuis, yeah i gounf os.ArgIteratorPosix

00:27 <scientes> but i just ended up using os.ArgIteratorPosix.raw

00:27 <tiehuis> that works too if you only need posix support

01:44 <kristate> andrewrk: it merged??

01:49 <kristate> andrewrk: when I saw in 598e809 that you changed from initMultiThreaded to initSingleThreaded, I figured that you had found #1363 -- bummer

03:00 <scientes> chaCha20(input[0..], key, nonce, &result);

03:00 <scientes> home/shawn/git/zig/std/crypto/chacha20.zig:138:19: error: expected type '[]u8', found '[]const u8'

03:00 <scientes> why is this a problem?

03:29 <scientes> how do i get my function to write to a slice buffer provided by the caller?

03:29 <scientes> pub fn chaCha20(in: []const u8, key: [8]u32, nonce: [8]u8, out: *[]u8) void {

03:30 <scientes> home/shawn/git/zig/std/crypto/chacha20.zig:102:16: error: index of single-item pointer

03:30 <scientes> otherwise I got my chaCha20 working

03:34 <scientes> bingo

06:44 <MajorLag> ..wait, what? Since when can return type be var? This changes everything.

06:54 <MajorLag> Oh, no wait. It still doesn't. I'm just too tired to compile the right file.

10:49 <Barabas> Howdy

10:50 <Barabas> I made a simplex noise implementation, and I get the values I expect (compared to the C++ implementation) in debug mode, but when I compile it with release-fast I don't. Any idea why this might be?

13:13 <Barabas> I hacked the compiler source code to always turn off fast math and now it's fine.

13:16 <Barabas> ... now I reverted it and it's still fine. wtf.

13:17 <Barabas> Oh wait... not building with release-fast ><

13:24 <Barabas> Ok, so it's really 'optimized' floating point mode which screws up my numbers. Most of the time it's fine, sometimes it's 2% off, depending on whether I do the calculations with 1 and 2 or 4 and more elements at the same time.

14:44 <hooo> weird ERROR: https://tio.run/##ZYzBCoJAFEX3fsVlNs1ASLUpEqNlfYO1GHXSAZ2JcSxL/XbTJFC6iwePc899y6TrIq0Ki8LG8HGU@V0bS0n/EuY5I3tyo/6gG4uwTNyBec69DHFTyLlUlOGhZYzawZAHN6heewTXcavc9UvkdCbeyG/agFavYOW6mytDI5uf@c0wT0ndXhRZQjJvgoqntFEKKtnM2K7hH@YeSbRFJqwVBot0Qdhy0hdZIQajVEbwKOVhJqa4dcbbdt0H

14:47 <MajorLag> what's weird about it?

14:47 <MajorLag> 'H' == 72

14:48 <hooo> it doesnt tell me whats wrong

14:48 <MajorLag> you reached unreachable code

14:48 <MajorLag> which is explicitly what you told it to do if you didn't match in the switch

14:48 <hooo> yeah but so what, this is an error?

14:48 <MajorLag> it's an assert

14:49 <MajorLag> you've told the compiler "this should never happen", but because you're compiling in debug, it catches the fact that it did happen, and panics with a line number of where it happened.

14:49 <hooo> well how do I do nothing in the else case?

14:49 <MajorLag> else => {}

19:01 <Dirkson> andrewrk: Did you ever consider that "cast(FromType, ToType, Castable)" suggestion of mine? I'm curious what you think of it, since it appears to sidestep the problems of typecasting via "SomeType(Castable)", while simplifying syntax.

19:18 <scientes> he's on vacation for a week

19:19 <scientes> Dirkson, <andrewrk> also, I'm going on vacation tonight. spending 1 week with my family. I might have time to merge PRs and respond to issues, but not much else.

22:15 <Dirkson> Ok, so zig errors when I attempt to not use an expression value. I assume this is to force people to deal with return values, for things like errors. But if I stuff the result into a value, then never use the value, the error vanishes. It seems like this could be fairly easily detected using static analysis. Is this "TODO" behavior, or am I misunderstanding something?

22:16 <Dirkson> It appears that local values are implicitly typed. Are there other things that are implicitly typed? (function return types, argument types, etc.)

22:21 <MajorLag> Dirkson, zig forces you to acknoledge an expression value, not necessarly do anything with it. We use `_` to explicitly ignore the value. `_ = func();`. Type is inferred under most circumstances, not not function returns (yet). For arguments we have `var`, but that's used for comptime duck-typing and generics.

22:27 <Dirkson> Oh. Is type inference planned for function returns?

22:30 <Dirkson> MajorLag: It seems to me that creating a value without using it should also be an error. (Or at least a warning, as it is in C)

22:40 <MajorLag> Zig doesn't have warnings. Either something is an error or it isn't. It's part of the "being explicit about intent" philosophy.

22:40 <Dirkson> Sure. Not really my point.