05:38 <IntoxicatedHippo> How do I create an array of function pointers?

05:39 <IntoxicatedHippo> Nevermind, I just realised I was doing `fn (type.type) void` instead of `fn (type) void`

06:24 <andrewrk> emekoi, that could mean that your union's memory got corrupted

06:24 <andrewrk> and you found out when you switched on it

06:30 <MajorLag> andrewrk: it looks like a struct can have a field 'a: fn() void' and an fn 'a(self: *Self) void'. 'self.a()' would seem to call the latter with the former seemingly being inaccessible. That should probably be a compile error, right? I'm about to go to bed but will submit an issue in the morning.

14:39 <bodie_> I'm trying to understand the call graph for resolve_struct_type for issue #1832 (https://github.com/ziglang/zig/issues/1832). field->type_entry is nullptr on a recursive type where the struct has an embedded arraylist of itself.

14:41 <bodie_> sorry resolve_struct_alignment not resolve_struct_type

14:46 <bodie_> I guess this happens because the field has a self-referential type which hasn't been fully resolved by this point.

14:50 <ross_k> Might be a dumb question, but how do I do a for loop over a numerical range in zig?

14:51 <ross_k> I see there are ranges for slices: [0..n] and switch statements: 0...n. Is there something like that to use with loops?

15:34 <bodie_> @andrewrk, is there any plan for supporting C++ libraries as a dependency?

15:35 <andrewrk> bodie_, no, no plans for that. that has language semantic requirements that zig can't do

17:14 <andrewrk> ross_k, use a while loop

17:14 <andrewrk> var i: usize = 0; while (i < 10) : (i += 1) {}

17:16 <ross_k> I see! Thanks.

19:36 <emekoi> can we get `and` to work on nullable and boolean types? so you can do `if (true_or_false and maybe_null) |not_null| {}` or `while (true_or_false and maybe_null) |not_null| {}`?

19:41 <andrewrk> emekoi, probably not

19:41 <andrewrk> these are the kind of things where we sacrifice convenience to maintain simplicity

20:13 <knebulae> Are there any suggestions for guidance on producing a UEFI application executable, linked with msvc .lib files (EDK-II)?

20:24 <knebulae> By way of background, these executables are just Win64-PE files with a different subsystem linked against the UEFI libraries rather than MSVCRT.

20:52 <andrewrk> knebulae, i've never tried that before, but it sounds like a use case we should think about and support

20:53 <andrewrk> knebulae, it might just work. if you don't use --library c then zig does not link against msvcrt

20:54 <andrewrk> we also have -mconsole and -mwindows. which subsystem are you talking about?

21:27 <daurnimator> andrewrk: interesting that you say that about C++.

21:27 <daurnimator> andrewrk: I was actually thinking of the related problem of using Qt from zig this morning

21:28 <daurnimator> Qt of course requiring a custom two pass C++ compiler...

22:01 <daurnimator> (moc http://doc.qt.io/qt-5/moc.html )

22:46 <nbjoerg> to be fair, a lot of that is historic garbage...

23:41 <daurnimator> As a *user*, Qt based programs are the best amoung the portable GUI toolkits.

23:41 <daurnimator> as a developer, qt is a huge pain in the arse.

23:54 <zachcarter> for retained mode x-platform GUI solutions, there are very few choices

23:55 <zachcarter> if your needs are simple enough - https://github.com/zserge/webview - might be of interest