Intent to Deprecate and Remove: AppBannerPromptResult interface

[Matt Giuca]

Primary eng (and PM) emails

mgi...@chromium.org

Summary

A very minor but web-facing change: I am converting AppBannerPromptResult from an interface to a dictionary. It will work exactly the same, but the name AppBannerPromptResult will be removed from the global namespace. The AppBannerPromptResult interface object itself is not useful.

 

Motivation

This should've been a dictionary in the first place (https://crbug.com/715400) since it's just a container for a bunch of return values. This is non-standard. The standard is slightly different (we are working towards matching the standard) and the equivalent concept is PromptResponseObject, which is a dictionary. The name discrepancy will be undetectable once it is a dictionary.

 

Interoperability and Compatibility Risk

Compatibility: This is technically a breaking change, since AppBannerPromptResult will no longer be visible in the global namespace. However, there is no practical reason why anybody would've been relying on this interface. You can't call the constructor so the only way to use this interface is to look at the one in BeforeInstallPromptEvent.userChoice (and that will not change). Compatibility risk is low.

Interoperability: This improves interoperability, since no other web engines include this and it isn't likely to be standardised. It is a dictionary in the standard.

Edge: Not supported

Firefox: Not supported, positive to removal

Safari: Not supported

Alternative implementation suggestion for web developers

It never had any use.

Usage information from UseCounter

No instrumentation for direct inspection of the AppBannerPromptResult interface object. I can't imagine anybody is using it. It doesn't do anything.

OWP launch tracking bug

https://crbug.com/715400

 

Entry on the feature dashboard

https://www.chromestatus.com/features/6560913322672128

[TAMURA, Kent]

LGTM1.

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAHqYdcYa4hSmPgsGKC6QSi8KwuY1KP3Zo0HVb-sLr%3D5vBBeATQ%40mail.gmail.com.

--

TAMURA Kent
Software Engineer, Google

[Jochen Eisinger]

lgtm2

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAGH7WqHVgaHTvjdwM6B_k1vfuVddWU8xSDvV5nHv8TV_MPM6rw%40mail.gmail.com.

[Philip Jägenstedt]

LGTM3. I looked in httparchive:har.2017_05_15_chrome_requests_bodies and while there were 800+ hits, none are real uses. If you do that kind of check yourself, I think it's fine to skip the intents process altogether.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CALjhuifcZVwPTPy9fbMNppOPv4zkWUUKnUVd4kbsFPRiAoa0-g%40mail.gmail.com.

[Matt Giuca]

Thanks,

I ran the query myself and found exactly 4 distinct JavaScript files (the 800 results are mostly all the same one), all of which just seem to have a list of everything in the Window namespace (not actually relying on the functionality).