Editorial changes. #120
Editorial changes. #120
Conversation
reference/src/layout/pointers.md
Outdated
| @@ -40,23 +39,23 @@ word and have therefore the same layout as C pointers. | |||
| > or, in the case of `&mut T`, aliasing. | |||
|
|
|||
| We do not make any guarantees about the layout of | |||
| multi-trait objects `&(dyn T + U)` or references to other dynamically sized types, | |||
| multi-trait objects `&(dyn Tr + Ur)` or references to other dynamically sized types, | |||
There was a problem hiding this comment.
Maybe use Trait0 and Trait1 here ?
There was a problem hiding this comment.
Changed to Trait1 and Trait2.
Where is this guaranteed ? (Can you point to an RFC, the reference, etc.?) |
Hmm, i never questioned this! No, i don't think i've seen any material explicitly said so. I think it is important for this to become a guarantee, or i won't be able to transmute a |
|
We'll discuss it in the next meeting (next thursday), this will give time to more people to take a look at this. |
|
I think it's fine to merge this as long as it doesn't make new guarantees -- and I don't think it does. :) See my one remaining open comment, though. |
Co-Authored-By: crlf0710 <[email protected]>
|
Thanks! |
We do not guarantee the layout of trait objects in any shape or form on stable. As for |
I'm not happy with several wordings within the text, so i did some small modifications.
Beyond this, i think
&dyn Trait + <auto trait>s + <lifetime bounds>are actually guaranteed, but this text didn't cover it somehow. And i don't really know the correct way to write them out. Besides this, i'm still puzzled about things like the validity of&dyn Send + Syncand&dyn Sync + Send...