Created attachment 8360678 [details] Contex_Menu.mpeg Description: While composing a new email, top left buttons are selectable behind the context menu. Repro Steps: 1) Updated Buri to BuildID: 20140113004002 2) Proceed to the Email app 3) Add an Email account 4) Compose a new email 5) Type anything in the body of the Email 6) Tap the back button twice 7) Press cancel button Actual: The button will be selected and open a new composing email. Expected: Buttons behind the context menu are not selectable until the context menu is closed. Environmental Variables: Device: Buri 1.3 MOZ BuildID: 20140113004002 Gaia: b3fc4f712562ee92b0ed0bd17abc61be9a36a8da Gecko: 5bb1837de7c0 Version: 28.0a2 Firmware Version: V1.2-device.cfg
Created attachment 8360681 [details] log.txt Does not reproduce on 1.2, I have attached a logcat. Environmental Variables: Device: Buri 1.2 COM BuildID: 20140102004001 Gaia: b1bc88386c781148a25091bf2eeee3ba217281d0 Gecko: 0c11156c7d9b Version: 26.0 RIL Version: 01.02.00.019.102 Firmware Version: V1.2-device.cfg
Josh - The bug is unclear here. I don't see what the problem here & why this is a bug. Can you clarify?
Definitely a bug, but not exactly what I think is being described. The log excerpt of note is this: 01-15 13:39:16.609: I/GeckoDump(3268): LOG: compose: back: save needed, prompting 01-15 13:39:17.779: I/GeckoDump(3268): LOG: pushCard for type: compose 01-15 13:39:21.479: I/GeckoDump(3268): LOG: compose: canceled compose exit 01-15 13:39:23.919: I/GeckoDump(3268): LOG: compose: back: no save needed, exiting without prompt The thing is that there is no button on a compose card that can cause a compose card to be pushed. There is a button in the upper right corner of the message list, and there are buttons on the message reader (of which there are none). So Josh is managing to click on the compose button that's off the edge of the screen. Probably because of touch region inflation and the fact that our overlay does not have anything that wants to process the tap in its upper left. This suggests we need to be making off-screen cards that aren't trays touch-insensitive or file a platform bug about touch inflation being buggy. Or other things. :jrburke, thoughts?
Okay - seems like an edge case (though a bug), so probably won't block on this.
Created attachment 8360877 [details] [review] GitHub pull request Remove the old hack for bug 828266, because event fluffing makes it possible to click on the "hidden" card areas now, and the hack is no longer needed with more recent gecko versions, no more black sections during card transitions. Asking :asuth for review since he has history with the old bug. Good-bye lovely hack!
Comment on attachment 8360877 [details] [review] GitHub pull request That would explain it! Totally forgot about the 0.5px!
Merged in Gaia master: https://github.com/mozilla-b2g/gaia/commit/4b6c34b7927daeaef5b3454ac54dcb5593cebedf from pull request: https://github.com/mozilla-b2g/gaia/pull/15391
Created attachment 8361965 [details] TouchthroughonMovelist.txt Found another issue that appears to be related to this one, so instead of making a new bug I will post the steps and logcat here. (If a new bug should be filed, I will do so. Apologies for comment spam.) Repro Steps: 1) Updated Buri to Build ID: 20140113004002. 2) Open E-Mail Application. 3) Go to Inbox in E-Mail App. 4) Select an E-Mail (any email) and attempt to move it to a different folder. 5) On the list that shows the available folders to move to, attempt to scroll up and down. Actual: It's possible to scroll through the Inbox in the background. Expected: Not able to scroll the inbox in the background. Environmental Variables Device: Buri v 1.3.0 Mozilla Build ID: 20140113004002 Gecko: http://hg.mozilla.org/releases/mozilla-aurora/rev/5bb1837de7c0 Gaia: b3fc4f712562ee92b0ed0bd17abc61be9a36a8da Platform Version: 28.0a2 Firmware Version: V1.2-device.cfg Notes: Repro frequency: 5/5 See attached: logcat (Dev options --> "Enable APZ for all Content.", "Enable Hardware Component", "Remote Debugging", and "Console Enabled."
That behavior should also be fixed by this changeset.