Intent to Ship: [MimeTypeArray | PluginArray | Plugin] # getter item() and namedItem() argument should be non-optional

[Luna Lu]

Contact emails

lun...@chromium.org

Spec

https://html.spec.whatwg.org/multipage/webappapis.html#mimetypearray https://html.spec.whatwg.org/multipage/webappapis.html#plugins-2

Summary

According to the spec, those getters' argument should not be optional. An 'index' or 'name' must be specified in order to get the item wanted in the array.

Is this feature supported on all six Blink platforms (Windows, Mac, Linux, Chrome OS, Android, and Android WebView)?

No

Interoperability and Compatibility Risk

Edge: Shipped

Firefox: Shipped

Safari: No signals

Web developers: No signals

OWP launch tracking bug

https://bugs.chromium.org/p/chromium/issues/detail?id=701560

Entry on the feature dashboard

https://www.chromestatus.com/feature/5763883465703424

[foo...@google.com]

lgtm1

[PhistucK]

Any use counter as this will start throwing, if I understand correctly?

☆PhistucK

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.

[Philip Jägenstedt]

There are no use counters for this, and IMHO adding them would be excessive since these arguments are already non-optional in Edge and Firefox. (I spot checked a few cases in each.)

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

[Luna]

I confirm that arguments for these APIs are non-optional. I verified them using BrowserStack. 

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.

[Luna]

Also, I'd also like to change the return type to be nullable, which also matches with the spec.

[TAMURA, Kent]

A similar change for CSSStyleSheet.insertRule() broke many sites, and we needed to revert it in a stable branch.

https://crbug.com/319695

Can you investigate their typical usage with http archive?

--

TAMURA Kent
Software Engineer, Google

[Rick Byers]

Yeah it's a good question whether more analysis is needed to confirm up front that WebKit-specific code depending on this is rare.

A key difference I see here between this issue and issue 319695 is that these APIs are quite niche, while CSSStyleSheet#insertRule seems likely to be more popular.  Also it's easier to imagine developers omitting the index argument ("insertRule(rule)") than it is to imagine how they would omit an item argument ("mimeTypes[]" is a syntax error, "mimeTypes.item()" is a pretty strange thing to imagine someone typing).  An HTTP Archive search is going to be tricky to have much confidence in (which obscure patterns in particular should we look for?  Do we think any would be more likely than the ones we can't find statically like "var o = navigator.mimeTypes; ... ; o.item();"?).  If we want real confidence here we probably need to add a UseCounter via custom bindings and wait for it to hit stable.

Personally my instinct is that the risk here is very low and we're better off just fixing low-risk cases like this and being on the look-out for regressions (err on a side of revert at the first report of an issue in the wild) than making them 10x more expensive to fix by requiring UseCounters for them.  We've got a long list of little low-risk changes like this, where the result of being more cautious is that we'll just attempt to fix much fewer of them.  So I'll give my LGTM2.  But this is definitely open to debate / judgement.  Other opinions?

Rick

[Jochen Eisinger]

given that Edge and Firefox have this shipped, I agree that it seems low risk, so lgtm3

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.