This change seems to have introduced a bug whereby you get an immediate keyup event for each key, right after you press the key. This leads to a number of issues where the system checks for the keystate for a 'shift' or 'control' key and always finds them in the keyup state.
r14129 is the revision with the problem: ------------------------------------------------------------------------ r14129 | ptw | 2009-06-12 17:58:00 -0400 (Fri, 12 Jun 2009) | 36 lines Merged revisions 14112 via svnmerge from http://svn.openlaszlo.org/openlaszlo/branches/4.2 ........ r14112 | ptw | 2009-06-10 17:01:41 -0400 (Wed, 10 Jun 2009) | 30 lines Change 20090609-ptw-V by [email protected] on 2009-06-09 16:28:34 EDT in /Users/ptw/OpenLaszlo/4.2 for http://svn.openlaszlo.org/openlaszlo/branches/4.2 Summary: Clear stuck keys on browser focus Bugs Fixed: LPP-8056 Alt tab issue when switching tasks (partial) Technical Reviewer: max (Message-ID: < [email protected]>) QA Reviewer: [email protected] (pending) Details: This fixes the bug in all browsers _except_ IE, where it only reduces the bug somewhat. I'm putting in this partial fix just so the work is not lost. LzKeyboardKernel.*: Add an __allKeysUp call that will simulate key up events for any keys that the kernel believes to (still) be down. LzKeys: __allKeysUp exported so it can be called from lz.embed embednew: When the browser regains focus, invoke lz.Keys.__allKeysUp to clear any keys that may appear to still be down because the browser lost focus due to keyboard input. Tests: Test case from the bug report ........ -- Henry Minsky Software Architect [email protected]
