Remove support for UC Browser

Is there a policy somewhere that specifies which browsers are targeted by Drupal core's CSS and JS?

@bradjones1 Is this https://www.drupal.org/docs/system-requirements/browser-requirements what you are looking for?

That would be the artifact of the decision, yes - I suppose the real underlying question is, who makes that list and how is it updated? E.g., this then becomes a [Policy, no patch] type issue, that would then spawn the code change in question?

I think this is a "policy, no patch" issue and should be renamed. Example: #3155358: [policy, no patch] Drop IE11 support from Drupal 10.0.x

Given that UC is not kept up to date (if it was, and is based on Chromium, we'd probably support it by virtue of the fact that we are supporting chrome on Android rather than by any extra effort) and given the security concerns raised, and given the tiny browser market share it has, I think I'm in agreement about dropping support for this.

@bradjones1 there is an open issue to define a policy for removing and adding browsers to the support list here: #3080068: [policy, no patch] Define usage heuristics for browser support. Until that's resolved, we have to do individual issues like this one.

If there's no indication that UC browser ever updates the browser engine, that's a compelling reason to remove support. But looked at usage too to get an idea.

UC browser was removed from android app stores in China last year - not clear if that was permanent, but if it was, that's likely going to kill it off entirely soon.

https://www.cnbc.com/2021/03/16/alibabas-uc-browser-deleted-from-android...

According to https://gs.statcounter.com/browser-market-share

Internationally: 1.45% in December 2020, 0.94% in December 2021 - so under 1% usage and on a downward trajectory.

Indonesia: 2.2% December 2020 to 1.45% December 2021 - negligible.

China: 19% in March 2021 (peak) to 11% December 2021 - not negligible but low and going down.

India: 5.43% Jan 2021 to 2.33% December 2021 - negligible.

So Indonesia it was not that popular last year anyway. India, usage halved down to < 2.5%. China, usage has halved in 9 months.

Looks like it's going to be more or less gone by the time Drupal 10 is released or not long after, so I think we should just drop it in 10.x.

I tested UC browser (playstore ver).

I believe this is an embed webview because their user agent is TOTALLY the same as my Android webview version.

That means we need to make decisions based on the Android System version (Android Browser??) rather than UC.

And some random research showed the same result:
https://user-agents.net/browsers/uc-browser/versions/13-0

If it's webview, I think we need a separate issue to define if/how we support that as a 'browser' in its own right. But that probably pushes this issue even more in the direction of dropping explicit support for UC browser.

I don't want to dismiss its market share in China, but as @catch is suggesting it'll probably go down over time. IMHO the cons are bigger than the pros.

On a front-end perspective a browser that doesn't give support to new important features (and won't give it in the future) can really slow us down and affect Drupal competitiveness on several ways.

So +1 for removing support for D10.

From what I found in #11, I think we need to fix it in the next release (9.4?)

The upstream data is incorrect I believe (just someone using a particular Android OS version to contribute the data)
So fixing in https://github.com/browserslist/caniuse-lite is another way to disconnect it.

If we don't totally drop it in 9.4, then I think we need to find the closest Chrome version and add it to our list
https://caniuse.com/ciu/comparison
(this page isn't 100% correct IMO, some features are missing)

** Note: I can't find a way to test UC 12.12 listed on caniuse.com. This is keeping crashed! I've tried diff Android OS versions already :s)

This one is tricky. It's still the second-most-used browser in China, and the initial decline after it was removed from the App store seems to have leveled off:

(Source: https://gs.statcounter.com/browser-market-share/all/china/#monthly-20200...)

On the other hand, this seems pretty clearly like an extenuating circumstance to me based on the IS.

#11 sounds like the right approach to me, although I'm not sure how to square it with our current policy.

This issue is "Policy, no patch", but lists browser support entries in package.json. Should a different issue be opened to change the package.json file or should this issue's title be changed?

We just discussed this with @lauriii and @bnjmnm and the 3 of us agreed on removing support for UC Browser. Removing the Needs frontend framework manager review tag.

Assuming this is ready for review, I've looked over patch-17/comment-18 and the changes are very straight forward. As far as I can tell, the only adjustment to the compiled assets is replacing self with globalThis. Does this need any additional review?

To meet the deadline, the main thing is updating https://www.drupal.org/docs/system-requirements/browser-requirements to add 'Drupal 9 and below only' to UC browser I think.

Tagging for a change record too.

Having said that this is essentially a one line patch + compiled assets, so it looks like an easy commit to me.

Also since this has consensus among the front end framework managers and +1s from both me and @xjm, I think it's OK to make the title just for implementation now.

I get the same result as #18 when rebuilding all CSS and JS without UC Browser in the browserslist. There are no remaining references to "UC Browser" or "UCAndroid" or variants that I can find in the codebase.

Tried to add a change record at https://www.drupal.org/node/3285017 but I don't know how to explain the difference between the version we are dropping support for and the WebView version which should still continue to work.

I think the change record actually explains the extremely confusing situation pretty well.

Committed/pushed to 10.0.x, thanks!

I'll update the docs page now too.