Tracking Issue for Read::is_read_vectored/Write::is_write_vectored.

[sfackler]

Unresolved Questions

• What's the right naming for these methods?

Implementation history

• Add Read/Write::can_read/write_vectored #67841

[workingjubilee]

Is there a situation under which this would not be better off as an associated constant?

[sfackler]

Those are not compatible with trait objects.

[bhgomes]

Those are not compatible with trait objects.

Is this a design choice, or just something that has not been implemented yet?

[sfackler]

That is inherent to how constants work.

[m-ou-se]

Is there a situation under which this would not be better off as an associated constant?

#105917 is an interesting case to think about: io::Chain. When chaining one stream that is write vectored with another one that isn't, io::Chain::is_write_vectored could make its return value depend on whether it's currently in the first or second stream, I suppose. (Not saying that it should. But we should probably document what is expected.)

[xiaoyawei]

Is there a timeline to stabilize this feature? Knowing whether a socket supports vectored I/O or not gives applications a bunch of optimization opportunities :)