Not approved. I feel that this fix is a layer of bubblegum on top of the baling twine that was used hold LPP-8445.
>From my review of your change for LPP-8445: > On 2009-09-10, at 07:50, P T Withington wrote: > >> Not approved. >> >> I don't understand the change on line 1300. I thought IE did not have >> mouseover/mouseout events? >> >> I feel the logic here is more complex than necessary. When you mouseup over >> a different sprite than the one you went down on, isn't the event target the >> one you went up over? In which case, it would seem the logic just needs to >> be: "if I get a mouseup on a sprite different from the one where I got >> mousedown, send mouseup to the sprite I went down on [and a >> mouseupoutside?], and send mouseover to the sprite I am over now". Won't >> that work, simplify the logic, and remove the need for __lastMouseOver and >> the additional global mouse handling? >> >> It seems you should either have one more flag to __findParents, or have >> another API, say __anyParents, that tests just for the existence of a parent >> with the property you are searching for, for the __isMouseOver test. It >> makes no sense to look for all the parents that are not visible, when just >> finding one is a sufficient test. We checked in the LPP-8445 check as an "emergency fix" because it was blocking wt. You filed http://jira.openlaszlo.org/jira/browse/LPP-8470 but we've never done anything about it, and I feel it has now come back to bite us. It really seems that the IE logic is getting more and more twisted here. Is there any way we can simplify this? In particular, it seems that at line 1599 we say "if we are in ie and we were going to delay sending mouseover, don't", but then at line 1610, we still set __lastMouseover, which you then have to ignore at line 1680! We need to take a fresh look at this. On 2010-06-28, at 15:50, Max Carlson wrote: > Change 20100628-maxcarlson-F by maxcarl...@friendly on 2010-06-28 12:44:18 PDT > in /Users/maxcarlson/openlaszlo/trunk-clean > for http://svn.openlaszlo.org/openlaszlo/trunk > > Summary: Fix spurious mouseover events in IE DHTML > > Bugs Fixed: LPP-9107 - DHTML - IE7 - Amazon - When drag the album, the frame > of album will be stay at original place > > Technical Reviewer: hminsky > QA Reviewer: ptw > > Details: IE needs an extra check to make sure the mosue is still over > __lastMouseOver before sending an onmosueover event. > > Tests: See LPP-9107, testcase at LPP-8445 runs as before > > Files: > M WEB-INF/lps/lfc/kernel/dhtml/LzSprite.js > > Changeset: > http://svn.openlaszlo.org/openlaszlo/patches/20100628-maxcarlson-F.tar
