00:00 <forgot-password> I'm using @memberName

00:00 <forgot-password> Or can I just append c""?

00:01 <daurnimator> `++ c""` should work actually

00:03 <forgot-password> Hm, sadly it did not. Does c"" get optmized away?

00:09 <daurnimator> andrewrk: is there a way to specify the field that contains the type for an enum?

00:09 <daurnimator> s/enum/union/

00:10 <daurnimator> I'm looking at our sockaddr definitions.... they don't seem right

00:21 <daurnimator> andrewrk: I'm thinking e.g. you could provide some 'special' method on a extern union?

00:25 <daurnimator> `sockaddr_storage = extern union { in: sockaddr_in, in6: sockaddr_in6, fn @unionEnum(self: sockaddr_storage) sa_family_t { return @intToEnum(sa_family_t, self.in.family) } }`

00:25 <daurnimator> (yes I know sa_family_t is not currently an enum; but it probably should be a packed enum?)

03:35 <emekoi> does zig link to libunwind?

03:38 <emekoi> because the repo includes a copy of the headers.

04:36 <andrewrk> emekoi, no it doesn't. where do you see that?

04:44 <emekoi> never mind, it's just the header unwind.h in c_headers.

14:16 <forgot-password> ilcear

14:18 <forgot-password> Is it possible to have a comptime value that I can manipulate at runtime? In my current attempt the compiler says 'cannot store runtime value in comptime variable'

15:39 <Pursche01> Hi, I just found out about Zig 30 minutes ago and I have been digging through the documenation a bit. It looks like an interesting language and I think I would like to give it a try. Before I do though, there is a concept I haven't found in the documentation. Does Zig have "pure" functions? Can I at compiletime force a function not to have side effects and not access non-const global scope?

15:40 <Pursche01> Alternatively are there any good tools for parsing Zig (preferably from Zig rather than from C), building an AST and figuring out "pureness" from a custom built pre-compiler?

15:41 <forgot-password> I'd love to have something like this too (I'm refering to the pure functions)

17:12 <MajorLag> the standard library has a zig parser in it

17:24 <forgot-password> Is it possible to do something along the lines of this: https://pastebin.com/5F0HyMKt?

17:25 <forgot-password> It already feels like fighting the language, but I'd love to know if it is plausible.

17:43 <MajorLag> the problem with what you've written is that someFunction's first parameter needs to be FirstStruct, but you've passed it SecondStruct. `second.someFunction()` is equivelent to `SecondStruct.someFunction(second)`. There is an implementation of interfaces in std that does what I think you're trying to do, however we're planning to move away from it because it doesn't optimize well. see: https://github.com/ziglang/zi

17:43 <MajorLag> interface pattern can be seen in `std.mem` and `std.heap` where it is used with Allocator and its implementations.

17:45 <MajorLag> Bascially the interface struct instance lives as a member of the parent struct and itself contains function pointers to the implementations in the parent, but those parent functions take a pointer to the interface struct and use @fieldParentPtr to get their own instance from that instance.

17:47 <MajorLag> Something like this: https://pastebin.com/P3Y8VrWV

17:50 <forgot-password> MajorLag: Thanks, I think you messed up your link in the first message

17:53 <MajorLag> https://github.com/ziglang/zig/issues/1829

17:53 <MajorLag> worked fine for me?

17:55 <forgot-password> It looks like the message was cut off, maybe my irc client has a limit on message length.

17:55 <MajorLag> this was all that was after the link: The current interface pattern can be seen in `std.mem` and `std.heap` where it is used with Allocator and its implementations.

17:57 <forgot-password> What? How is that, it's displayed as a separate message for me picking up at " interface pattern[...]" again

17:58 <MajorLag> Your client might not like colons? What does: this look like?

17:58 <forgot-password> Heh, weird https://i.imgur.com/xpkkl4h.png

17:58 <forgot-password> It looks fine

17:59 <MajorLag> huh. weird

18:04 <forgot-password> I'm going to put together a little write-up of what I'm currently trying to do. I feel like I'm too focused on the current approach, so maybe anybody has an idea for a different implementation.

18:50 <Pursche01> @MajorLag: Do you think the standard library parser will be able to figure out if something has side effects?

19:03 <andrewrk> Pursche01, no, it just parses into an abstract syntax tree

19:03 <forgot-password> Pursche01: I suppose you could just (heh, "just") walk through the AST and check whether it does something that would classify it as having side effects. However, neither does that sound like a small undertaking nor do I know about the parser. So don't take my word for granted :p

19:03 <andrewrk> we used to have the concept of implicit pure functions

19:04 <forgot-password> Phew, I finally finished my small write-up: https://gist.github.com/schroffl/17150efeedf18f5ca3ef3a724344f247

19:04 <andrewrk> and if a function call had all comptime known arguments and was pure, then the function would be interpreted at compile time automatically

19:05 <andrewrk> Pursche01, the closest issue we have open to your question is https://github.com/ziglang/zig/issues/425. you are also welcome to open a proposal with your purity feature

19:05 <andrewrk> relevant to #425, it's possible that "pure" and "wanting to be comptime if possible" are the same thing

19:06 <forgot-password> andrewrk: So you ditched that for comptime expressions?

19:07 <andrewrk> I can't remember why I changed that. there are a few things to solve if we wanted to reintroduce it