Add [TreatReturnedNullStringAs] support for ByteString.

Add [TreatReturnedNullStringAs] support for ByteString.

If my understanding is correct, we need "[TreatReturnedNullStringAs=Null] ByteString get(ByteString key);" for Headers class to return null.
http://fetch.spec.whatwg.org/#dom-headers-get


BUG=347426
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=176370

[horo, 2014-06-17 14:08:02]

nbarth@
Could you please review this?

[haraken, 2014-06-17 14:10:35]

+jsbell

LGTM

[jsbell, 2014-06-17 19:00:30]

lgtm

Note that the spec gives the return type as ByteString? - i.e. nullable string.

We could do this by having the return type in C++ be Nullable<String>. We
support that for attributes; I'm not seeing explicit tests for that for method
return types, though, but it should "just work". Making bindings code use
Nullable<String> rather than Blink's special null strings seems like something
we should tackle later. So again, lgtm.

[haraken, 2014-06-17 23:50:11]

On 2014/06/17 19:00:30, jsbell wrote:
> lgtm
> 
> Note that the spec gives the return type as ByteString? - i.e. nullable
string.
> 
> We could do this by having the return type in C++ be Nullable<String>. We
> support that for attributes; I'm not seeing explicit tests for that for method
> return types, though, but it should "just work". Making bindings code use
> Nullable<String> rather than Blink's special null strings seems like something
> we should tackle later. So again, lgtm.

Whether we should introduce Nullable<String> or not is something we're
discussing in this long thread (https://codereview.chromium.org/312683005/). jl@
is going to make a proposal, so we might want to wait for that :)

Either way, this CL LGTM.

[horo, 2014-06-18 00:16:26]

The CQ bit was checked by horo@chromium.org

[commit-bot: I haz the power, 2014-06-18 00:16:40]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/horo@chromium.org/337373002/40001

[commit-bot: I haz the power, 2014-06-18 01:08:22]

Change committed as 176370

[Nils Barth (inactive), 2014-06-18 02:40:56]

On 2014/06/17 19:00:30, jsbell wrote:
> lgtm
> 
> Note that the spec gives the return type as ByteString? - i.e. nullable
string.
> 
> We could do this by having the return type in C++ be Nullable<String>. We
> support that for attributes; I'm not seeing explicit tests for that for method
> return types, though, but it should "just work". Making bindings code use
> Nullable<String> rather than Blink's special null strings seems like something
> we should tackle later. So again, lgtm.

To concur: we should support *returning* null as well as *passing in* null.
This may be easy (probably lots of CG work though), but it's currently not
supported.
...and sounds like Jens is looking into properly implementing this!