Phil, Good to know in case we need to make them match. I'm inclined to go with fewer events right now since problems on FF 3.6 may be exacerbated by not keeping up with events.
- Don On Jun 20, 2011, at 3:40 PM, [email protected] wrote: > Approved. > > The only oddity I saw in FF 3.6 is after you right-click on the view, a > mouseout event is never generated. When I tried it on IE, I see a mouseout > event after the context menu is displayed. > > > > On Mon, Jun 20, 2011 at 2:44 PM, Donald Anderson wrote: > > Change dda-20110620-3lu by [email protected] 2011-06-20 14:21:12 EDT > in /Users/dda/laszlo/src/svn/openlaszlo/trunk-g > for http://svn.openlaszlo.org/openlaszlo/trunk > > Summary: Remove events related to context menu > > New Features: > > Bugs Fixed: LPP-9967, LPP-9957 > > Technical Reviewer: philip (pending) > QA Reviewer: (pending) > Doc Reviewer: (pending) > > Documentation: > > Release Notes: > > Overview: > This addresses two issues, one related to LPP-9957: > the extra MOUSE_MOVE event generated immediately after MENU_SELECT > is now eliminated. > > And for LPP-9967, we now know that on FF3.6/Windows, > when a right click is done, there are two additional events > that occur *before* the MENU_SELECT event: > MOUSE_MOVE > MOUSE_OVER > Because they occur before the MENU_SELECT, there's no way to distinguish > these events from any other MOUSE_MOVE or MOUSE_OVER that might normally > occur. However, as indicated in Jira, occasionally these events have > bogus coordinates (usually stageY is about 150 pixels too big, sometimes > stageX is increased). We try to detect these errant events (in the > narrowest way possible, to avoid eliminating good events). As MOUSE_MOVE > or MOUSE_OVER occur, their X/Y values are saved (last two pairs are > saved), > so if a sequence of MOUSE_MOVE/MOUSE_OVER occurs with a X/Y that is 'far' > from the last one, it is silently dropped. The MOUSE_LEAVE that occurs > as a part of the overall sequence resets the saved X/Y values. > > It's not pretty, but it works. Frankly, I can't see a better solution, > except to require users to upgrade to newer FF. > > Unknown whether this will have any affect on browser hangs seen by > Maynard. I have not seen such hangs in my VM, so I can't tell. > > Details: > WEB-INF/lps/lfc/kernel/swf9/LzSprite.as > - change name of quirk to 'ignoreextramenuevents' to be more descriptive > of the current situation. > - narrow the quirk to only apply to FF3.6 on Windows (instead of all > FF3.6) > - remove any events that occur between MENU_SELECT and MOUSE_LEAVE > - call LzMouseKernel.ignoreMouseEvent to track/make decisions about > events with bogus coordinate. > > WEB-INF/lps/lfc/kernel/swf9/LzMouseKernel.as > - more internal doc to describe the messy bug and solution > - change name of quirk to 'ignoreextramenuevents' to be more descriptive > of the current situation. > - use __{prev,last}Stage{X,Y} variables to track coordinates > (since JS has no typed arrays, and we want speed here) > - ignoreMouseEvent tracks the previous two sets of coordinates, > and determines if the current event can be removed. > - use ignoreMouseEvents to remove events with bogus coordinates > when they happen. > > Tests: > > Files: > M WEB-INF/lps/lfc/kernel/swf9/LzMouseKernel.as > M WEB-INF/lps/lfc/kernel/swf9/LzSprite.as > > Changeset: http://svn.openlaszlo.org/openlaszlo/patches/dda-20110620-3lu.tar > > > > -- > > Don Anderson > Java/C/C++, Berkeley DB, systems consultant > > voice: 617-306-2057 > email: [email protected] > www: http://www.ddanderson.com > blog: http://libdb.wordpress.com > > > > -- Don Anderson Java/C/C++, Berkeley DB, systems consultant voice: 617-306-2057 email: [email protected] www: http://www.ddanderson.com blog: http://libdb.wordpress.com
