Should flags be represented as enums?

[xtian]

Regarding this TODO:

zigwin32gen/src/genzig.zig

Line 1669 in b42acec

try writer.line("// TODO: This Enum is marked as [Flags], what do I do with this?");

Is there a big benefit to representing these values as non-exhaustive enums? My impression is that the common case is combining multiple bitmask values together which is awkward with the current setup. For example:

const std = @import("std");
const win32 = @import("win32").everything;

const class = std.mem.zeroInit(win32.WNDCLASSEXA, .{
    .style = @intToEnum(win32.WNDCLASS_STYLES, @enumToInt(win32.CS_HREDRAW) | @enumToInt(win32.CS_VREDRAW) | @enumToInt(win32.CS_OWNDC)),
});

[marler8997]

Yeah I agree that in the case of enum flags, you always have to convert to an int and back to the enum which throws type safety out the window. I'd be curious if anyone has a better idea than just raw integer types though.

[marler8997]

Looks like this zig proposal could be a solution: ziglang/zig#5049

It updates packed structs to specify an underlying integer type so they are compatible with the C abi. If that proposal is going to be accepted, we might be able to start transitioning to it beforehand.

[marler8997]

Ok I came up with one idea, I've added an initFlags method to all enum flag types for now. So your example could become this:

const class = std.mem.zeroInit(win32.WNDCLASSEXA, .{
    .style = win32.WNDCLASS_STYLES.initFlags(.{.HREDRAW = 1, .VREDRAW = 1, .OWNDC = 1});
});

Full API diff for this change is here: marlersoft/zigwin32@89203c6#diff-bd4f828bddb85d1324457f27ad66278624ef2539c5036de4c596d3b15a695820R3458

It affected 11,500 lines!

[xtian]

@marler8997 Ok that seems like a decent solution for now. I like the idea of integer-backed packed structs though.

Here's at least one case that your solution doesn't handle:

VIRTUAL_ALLOCATION_TYPE is evidently not marked as "Flags", but VirtualAlloc can be called with MEM_COMMIT | MEM_RESERVE. VIRTUAL_ALLOCATION_TYPE also isn't a non-exhaustive enum, so the @intToEnum method doesn't work either.

[marler8997]

@xtian good catch. I've created an issue with the win32metadata project here: microsoft/win32metadata#507

[marler8997]

Just did a new zigwin32 release. Looks like VIRTUAL_ALLOCATION_TYPE is now properly marked with the [Flags] attribute (see https://github.com/marlersoft/zigwin32/blob/146598875015d7b1597af579ce0d41a93e1f4144/win32/system/memory.zig#L374).

[capthehacker99]

So how do you check if a flag exist?
Mentioned in this issue: marlersoft/zigwin32#28

[marler8997]

zigwin32 has moved on to using packed structs to represent flags which are very nice to use, so this issue should be addressed now.