Add Informative Example to canEncode

[JackStouffer]

I need this for my date string parsing library.

[DmitryOlshansky]

return s.representation.find!(x => !isASCII(x)).empty

[yazd]

or return s.representation.all!isASCII();

[9il]

Yes, @JackStouffer move to return s.representation.all!isASCII();. This will create less template bloat

[DmitryOlshansky]

Truth be told it's a hard sell:

• one liner
• wrong module, should be std.string
• trivial

[9il]

The bug is fixed now

[JackStouffer]

@DmitryOlshansky There were two separate bugs for this and a forum thread, so I think the demand is high enough for this. Fixed other issues.

@9il The bug is only deprecated, imports will still be public if you use global selective imports. From the change log:

To ease updating existing code, the old behavior was retained but deprecated.

[DmitryOlshansky]

LGTM then

[DmitryOlshansky]

Probably use the string instead of the array

[JackStouffer]

@DmitryOlshansky fixed

[DmitryOlshansky]

And I recall @andralex wants vet all name additions to the library.

[andralex]

I guess something more useful would be getEncoding which returns an enum value - one of ascii, utf8, binary.

[JackStouffer]

Hmm, my thoughts

1. If we go with that route, then this should use the encoding types of std.encoding to be DRY
2. So maybe a string overload of canEncode?

[WalterBright]

I'm not too thrilled about the utility of a one liner. It would be ok if it was a loop, but it's not. It's longer for the use to read the documentation than just write the one liner.

Phobos should not be a repository of one liners. Masses of one liner functions tend to hide the meatier functions. Phobos should make it possible for users to write one liners to get their work done.

[JackStouffer]

I changed the function to a helper overload of canEncode rather than a name edition.

[DmitryOlshansky]

That seems better. @WalterBright?

[WalterBright]

Yes, it is better. But I would rather see the one liner simply as an example in the documentation for canEncode(). It is better to teach people how to put lego blocks together than to provide trivial combinations of two lego blocks, or 3 lego blocks. It's the same reason there isn't a function to multiply x by 2.

My other complaint is the code:

s.representation.find!(x => !canEncode!E(x)).empty

is wrong. It should be:

s.byDchar.find!(x => !canEncode!E(x)).empty

The fact that nobody noticed this error suggests that we need more examples in the documentation on how to do these sorts of things. I would love to see more one liners like this as examples.

[JackStouffer]

Seems reasonable.

[JackStouffer]

Changed to example.

Perhaps 11356 should be marked WONTFIX?

[WalterBright]

Auto-merge toggled on

[WalterBright]

Changed to example.

Thank you. Looks good!

Perhaps 11356 should be marked WONTFIX?

Good idea. Done.

[JackStouffer]

Thanks!

[andralex]

Great. Thanks folks!