W3C home > Mailing lists > Public > public-script-coord@w3.org > July to September 2016

Re: Using objects for a mapping argument

From: Jeffrey Yasskin <jyasskin@chromium.org>
Date: Mon, 26 Sep 2016 15:03:52 -0700
Message-ID: <CANh-dX==iLtJhKmUo5R_LJia50-O3RqLthu31QESrzu8wi9zag@mail.gmail.com>
To: Boris Zbarsky <bzbarsky@mit.edu>
Cc: public-script-coord <public-script-coord@w3.org>
On Mon, Sep 26, 2016 at 11:41 AM, Boris Zbarsky <bzbarsky@mit.edu> wrote:
> On 9/23/16 9:23 PM, Jeffrey Yasskin wrote:
>>
>> In
>> https://api.csswg.org/bikeshed/?url=https://raw.githubusercontent.com/jyasskin/web-bluetooth-1/masked-data-prefixes/index.bs#device-discovery
>> defines some arguments to a function in which I want users to express
>> a map from integers or strings to a dictionary. One of the arguments
>> will look like {0x004C: {dataPrefix: [0x02, 0x15]}}, while the other
>> could be {0xFEAA: {dataPrefix: [0x10]}, "battery_service": {}}.
>
>
> For your case, where each value has the same "shape" (in that they are all
> converted to the same dictionary), this is basically similar to the
> OpenEndedDictionary proposal, right?
>
>> This is processed at
>>
>> https://api.csswg.org/bikeshed/?url=https://raw.githubusercontent.com/jyasskin/web-bluetooth-1/masked-data-prefixes/index.bs#ref-for-dom-bluetoothlescanfilterinit-manufacturerdata-5
>> using explicit calls to `.[[OwnPropertyKeys]]()` and some other
>> ECMAScript operations.
>
>
> Right, so compared to the other uses of OpenEndedDictionary this has the
> following additional wrinkles:
>
> 1)  Some constraints on the "keys" of the dictionary.  This is fine; other
> consumers want this too; it's simple enough to express in prose.

Yep, I'd be happy with an OpenEndedDictionary<BluetoothDataFilterInit>
that let me write "for each <var>key</var>/<var>value</var> pair in
the dictionary, do: ..." even if I had to check and convert the keys
from strings myself. To help with that conversion from strings, it'd
be nice to have a pointer from WebIDL to CanonicalNumericIndexString()
or the right string->number conversion function if that's not it.

> 2)  The way you set up properties on canonicalizedFilter.manufacturerData is
> using [[Set]] instead of [[DefineOwnProperty]].  You should probably use the
> latter.

Probably via CreateDataProperty(). Done.

> 3)  Since you're going around creating ES objects, there's the usual
> question of whether you want to use the global of the callee function or the
> global of the "this" value to do that.

Yes, creating and filling in ES objects from algorithms is difficult.
For dictionaries, WebIDL itself says "Let O be a new Object value
created as if by the expression ({}).", which doesn't answer the
global question either.

>> Does this make sense, or should I have my users express these
>> arguments in some other way?
>
>
> In general, this makes sense.  We should really get OpenEndedDictionary
> specced....

Is https://github.com/whatwg/fetch/issues/164 the main issue for that?

Jeffrey
Received on Monday, 26 September 2016 22:04:41 UTC

This archive was generated by hypermail 2.3.1 : Monday, 26 September 2016 22:04:42 UTC