Emilio Cobos Álvarez
Emilio Cobos Álvarez
FWIW, Servo's / Firefox's logic for computed ``: https://searchfox.org/mozilla-central/rev/403038737ba75af3842ba6b43b6e2fb47eb06609/servo/components/style/values/computed/position.rs#43 And specified: https://searchfox.org/mozilla-central/rev/403038737ba75af3842ba6b43b6e2fb47eb06609/servo/components/style/values/specified/position.rs#159
Changing how a selector matches based on computed styles is quite tricky. What if you set display: contents on an slotted selector?
I also had this issue with the GUI installer (with openJDK 8 on ubuntu 15.04). Cloning the repo and running `ant deploy` allowed me to install eclim.
Hmm, weird... Is some of the transformations here not valid for some alignment / not what https://github.com/rust-lang/rfcs/blob/master/text/2195-really-tagged-unions.md specifies? Or is it a rustc bug? I'd have to check carefully. cc...
Curious, what's the use case for such constructs?
Can you annotate them on the source with `/// cbindgen:ignore`?
Hmm, good point. This might generate compiling code with `associated_constants_in_body = true`, but yeah we need to track the dependencies of these as well to generate the right order, and...
I'm confused. If you have something like `#[repr(C)] struct WrId(usize);`, how could you avoid it in the header?
In the header that contains the function signatures, I mean. You could potentially split out the stuff that is only used by reference / pointer or what not. Maybe. But...