Dpad movement keys stop working

[severinrudie]

Steps to reproduce

(Mostly unknown. @liuche and I have both occasionally ran into this, but have not worked out any repro steps)

• Select tile

Expected behavior

• Cursor should be available

Actual behavior

• Cursor is not available
• Dpad movement clicks do not reach MainActivity#dispatchKeyEvent
• If the menu is opened, the user has no way of moving between elements

Device information

Samsung TV

• Fire TV device: ?
• Firefox version: ?

(I have been trying to reproduce this today, without any luck)

[severinrudie]

@npark-mozilla @rpappalax Have either of you encountered this one?

[severinrudie]

From @psymoon via Slack

that may be related to requestFocus race condition with sessions if you spam back and forward button, you sometimes get that take a look at onLoadingStateChanged and onNavigationStateChanged they confluence with actual UI update

[liuche]

@npark-mozilla can you help us try to repro this? It's a pretty bad state but we can't seem to get an STR.

[nojunpark]

@liuche I tried to repro this bug on basic stick, but I wasnt able to see it once (I did find a couple of unrelated issues during the process though). Is this a different issue from #1153 or #1146 ? (guessing one of them is a dupe)

[severinrudie]

@npark-mozilla #1153 and #1146 look like duplicates of each other, but this issue is distinct from both. Do you ever remember seeing this? It's rare, but it sounds like @liuche, @psymoon, and I have all run into it at least once.

[nojunpark]

@Baron-Severin I haven't run into this yet - although I didn't play much with firetv recently. I think I used to see something like this within the youtube app, and I somehow managed to keep the pointer circle alive within youtube, but that was it.

[Sdaswani]

@athomasmoz I'm moving this from being a release blocker, as it is hard to repro. I'll keep it in the milestone and we can decide to punt it out if needed when we code freeze.

[Sdaswani]

@npark-mozilla if you happen to find STRs do let us know.

[mcomella]

@Baron-Severin What are the chances that this is #1202 ?

Otherwise, next steps for me would be to audit what's changed from v2.2 to v3.0 (which is hard: a lot of it is in components), particularly looking out for focus changes and key event handling changes.

[mcomella]

If dupe of #1110 is correct, then we get these log statements when you press a button, even the ones that are working:

2018-09-12 17:11:18.786 3564-3842/? W/bt_btif: bta_hh_co_is_peer_amazon_remote: result = 1, vid = 0x1949, pid = 0x410
2018-09-12 17:11:18.786 3564-3842/? W/bt_btif: bta_hh_co_is_no_unsniff_amazon_remote: result = 1, vid = 0x1949, pid = 0x410
2018-09-12 17:11:18.898 3564-3842/? W/bt_btif: bta_hh_co_is_peer_amazon_remote: result = 1, vid = 0x1949, pid = 0x410
2018-09-12 17:11:18.898 3564-3842/? W/bt_btif: bta_hh_co_is_no_unsniff_amazon_remote: result = 1, vid = 0x1949, pid = 0x410
2018-09-12 17:11:18.900 3446-30185/? E/AmazonInputManagerService: Toggle bit button map has not been initialized!

[mcomella]

I wonder if this will be fixed by the fix for #1216.

[severinrudie]

I wonder if this will be fixed by the fix for #1216.

🤞

[mcomella]

Not actively working on this. Since we think it may be fixed with #1216, I think the best next steps would be to wait to see if it happens again before investing more time into this.