In D277411#9627719, @ohall wrote:

All android WebAuthn work requires manual verification.

Would this be worth filing a UI test for that could do end-to-end testing?

Although I check Cargo.lock in root, we disable a feature that bincode is used then no dependency for icu_provider. Is bincode detection incorrect when using workspace (See https://github.com/unicode-org/icu4x/issues/7196#issuecomment-3462449250)? Or do all new fix by updating any crates causes bincode warning?

Please submit the fix that is buildable. Also, you should fit with https://firefox-source-docs.mozilla.org/mobile/android/geckoview/contributor/geckoview-quick-start.html at first if an API is changed/deprecated. Also, some linters won't pass this fix.

r+ for GeckoView part.

This revision requires a Testing Policy Project Tag to be set before landing. Please apply one of testing-approved, testing-exception-unchanged, testing-exception-ui, testing-exception-elsewhere, testing-exception-other. Tip: this Firefox add-on makes it easy!

r+ for GeckoView change

This revision requires a Testing Policy Project Tag to be set before landing. Please apply one of testing-approved, testing-exception-unchanged, testing-exception-ui, testing-exception-elsewhere, testing-exception-other. Tip: this Firefox add-on makes it easy!

MOZ_GECKOVIEW is defined by set_define, so it might be unnecessary to DEFIENS["MOZ_GECKOVIEW"] in moz.build. I will file new issue as follow up.

In D277569#9611711, @robwu wrote:

... but I don't see a set_config call that sets MOZ_GECKOVIEW. It is possible that it is set somewhere else I am not familiar with, but I really want to confirm that. Otherwise following your suggestion would mean that the build would include unnecessary files and code due to #ifndef MOZ_GECKOVIEW no longer working.

The objective of the bug is to replace MOZ_FENNEC. From my perspective, it looks like that work has completed. Therefore I am approving the changes for extensions-code.

Although I am suggesting to use if CONFIG["MOZ_GECKOVEW"]: at first, why don't you use it?

MOZ_PLACE means that places history database that backend is sqlite. MOZ_GECKOVIEW_HISTORY is defined for transition from Fennec (old Firefox Android until version 68) to GeckoView. And GeckoView delegates history functions to browser application via GeckoView API, not using places database. So you should replace MOZ_GECKOVIEW_HISTORY with MOZ_GECKOVIEW simply.

MOZ_GECKOVIEW doesn't have any string. This is defined or not.

You should add the annotation to skip test_UpdateUtils_updatechannel.js for Android. (See my last comment)

https://searchfox.org/firefox-main/rev/b3c5c336322fca6129b9bae999ebaac61deb5eca/toolkit/modules/tests/xpcshell/xpcshell.toml#118 should have skip-if like the following at least.

And adding @masayuki.

@botond, I am not sure whether eMouseHitTest (MozMouseHittest) is necessary. According to the comment of DispatchHitTest, it isn't safe to remove it for touch and mouse interoperability. Do you know whether this removal is safe?

Could you use MOZ_GECKOVIEW instead of ANDROID.

Thanks, GMP is disabled on all Android (https://searchfox.org/firefox-main/rev/3d9230cb3475081d20465e3f84b733c572a94513/toolkit/modules/moz.build#252-259).

Thanks, this was for fakespot support.

testing-exception-other: disabled as temporary

Could you file a new issue to turn on this tests on Android 14? Or do you keep it as leave-open?