Set Webview back to true for certain css values.

[jpmedley]

Updates #4290 and #4291 by setting WebView back to true. Chromium WebView didn't arrive until version 37 and I have no way of finding out when or if pre-Chromium WebView supported this.

[ddbeck]

Hi @jpmedley, I have a question on this. When I was researching those features, I set the values that I did with the reasoning that the features existed in WebKit well before Android's first release. Apart from someone getting their hands on a G1 and testing it, what evidence would be good enough to set the value of a pre-37 WebView?

[jpmedley]

As far as I know there is no other way.

FWIW, I have previously made the argument that old Android WebView and Chromium WebView should be separated somehow. I'm not sure why that was shot down. Removing Android WebView completely was shot down on the basis that somebody somewhere might have a device that still has it. Since Android WebView has not been updated in five years, its usefulness for actual browsing is limited and shrinking by the day.

[ddbeck]

Thanks, Joe.

@Elchi3 can you shed any light on the history that Joe's talking about here? As it stands, it sounds like Joe's position is that we can't say anything about pre-37 WebView—there's no evidence good enough, short of an actual test. If that's the case, then we're going to have problems getting to real versions for loads of old features (I'd expect most of html.elements.* to have to be true for WebView under this standard, for instance).

[queengooborg]

I think the reason why the idea of separating pre-37 WebView and Chromium WebView as two separate browsers is for the same reasons why we haven't separated Opera out -- it's still the same browser, just with a different engine. I do recall the argument of removing pre-Chromium WebView entirely, and I'm still in support of it myself, since the chances of us obtaining said compatibility data is slim to none.

[Elchi3]

The story about our WebView versions is in #2690.

We've followed caniuse and what is outlined in the PR is also what MDN contributors have traditionally written down in the legacy wiki tables. So, it is data this project inherited and we've wanted to preserve it as it seems useful still (old Androids still being around). We probably might want to look at marketshare data instead of sharing feelings here, though.

Also, generally, the BCD project works closely with the vendors on how they want their data to be represented (for example, Microsoft decided now to drop the edge mobile data and we're doing that now, also as it is unmaintained data. So, we are indeed listening to vendors!).

So, if the vendor (Google) thinks it would be better to drop pre-37 Webview versions, then we could talk about that. I don't recall where this was being proposed, if it was, but I think we can have that discussion. So, no, @jpmedley, I don't think you or anyone was shot down. It is just that this wasn't put on any agenda to be discussed with the various stakeholders (caniuse, and BCD peers/owners, BCD data consumers [vs code, webhint]) officially. Feel free to start an issue and so we can get everyone's thoughts on that matter. Even if we would decide against dropping pre-37, it would be nice to have an issue for it where we have collected our reasons.

[chrisdavidmills]

@Elchi3 I do recall @jpmedley mentioning his thoughts on this to me, some time ago. If the onus was on me to bring it to the rest of the team, then it looks like I forgot, and for that, I apologise.

So yes, let's start a thread to get this discussed properly.

[Elchi3]

I think the last thing I heard from Joe was this comment: #3884 (comment)

[ddbeck]

@jpmedley @Elchi3 @chrisdavidmills I've opened #4330 to focus on the narrow question of what to do with pre-37 WebView data. Once we have a decision, I'd like to see us systematically resolve pre-37 WebView data; with that in mind, I'm closing this PR.