embrace RLS even harder #2131
embrace RLS even harder #2131
Conversation
src/DocumentStore.zig
Outdated
| @@ -184,7 +184,7 @@ pub const Handle = struct { | |||
| /// private field | |||
| impl: struct { | |||
| /// @bitCast from/to `Status` | |||
| status: std.atomic.Value(u32) = .init(@bitCast(Status{})), | |||
| status: std.atomic.Value(u32) = .init(0), | |||
There was a problem hiding this comment.
was a little bit iffy on this one, it arguably loses a bit of meaning but im not sure it matters too much?
There was a problem hiding this comment.
initalizing to zero assumes that Status{} is all zeroes. This may not be the case in the future so I would prefer the old code.
src/DocumentStore.zig
Outdated
| zoir_outdated: bool = undefined, | ||
| _: u23 = undefined, | ||
| zoir_outdated: bool = false, | ||
| _: u23 = 0, |
There was a problem hiding this comment.
not related to RLS at all but defaulting to undefined always feels really gross to me so this is more personal opinion than an objective change, fine with reverting it along with the above change though.
There was a problem hiding this comment.
This is done because the [zir;zoir]_outdated field should never be accessed before checking has_[zir;zoir]. I don't mind this change so feel free to keep it. What should be reverted however is the change from _: u23 = undefined, to _: u23 = 0,. This field should never be accessed so there is no need to initalize it.
There was a problem hiding this comment.
seems reasonable, part of why i changed these was so that i could change the .init in the other comment but no need anymore
xdBronch
force-pushed
the
push-prwwnulqkvuk
branch
from
3f8d836 to
ff35463
Compare
| }, | ||
| .child_declarations = .{ | ||
| .small = [_]Declaration.OptionalIndex{.none} ** Scope.ChildDeclarations.small_size, | ||
| .small = @splat(.none), |
There was a problem hiding this comment.
just wanna acknowledge how great this one is, thanks mlugg
No description provided.