std.traits: provide short alias for FunctionAttributes

[wilzbach]

It's quite common to abbreviate FunctionAttribute with FA - this just makes it official.
As you can see even all the tests use it.
The enum is only used for functionAttributes, e.g. functionAttributes!func & FA.pure_ which already explains the name.

Note: it doesn't try to change the existing name nor deprecate it.

General question: why don't we provide the abstraction hasFunctionAttributes!(func,FA.pure_)? It just seems to me that's what the user wants anyways and bitwise OR-ing seems not really like a Phobos-style.

edit: you even already provide isSafe and isUnsafe ...

[ntrel]

why don't we provide the abstraction hasFunctionAttributes!(func,FA.pure_)?

I think we should add it, but maybe taking strings instead:

hasFunctionAttributes!(func, "pure", "@safe")

[JackStouffer]

LGTM

[DmitryOlshansky]

I think we should add it, but maybe taking strings instead:

Agreed

[wilzbach]

Okay will send hasFunctionAttributes after dconf ;-)
Anything that needs to be added here? Should we add a notice somewhere?

[dnadlinger]

I disagree with adding the name. A local alias or with statement works just fine. This stuff is used infrequently enough (I say that having written parts of the related code) that having a two-character name is even more confusing to the end user, because they won't be familiar with it.

[dnadlinger]

Also, as a general thing, the alias would need to be documented (at least using /// ditto) if it's a public name.

[AndrejMitrovic]

Yeah I'm against this too, it just introduces a poorly named global name and will likely cause the dmd spellchecker to often suggest using this symbol.

[DmitryOlshansky]

Not very popular idea;) going to close it