I have noticed two other regression issues with the patch to "Store in the server all the window info... " (01/10/22 15:08:34). Both of these occur in Outlook Express 4.
1. SetTimer call is not always generating WM_TIMER messages as it should. I see:
0806bb20:Call user32.SetTimer(00020057,00001003,00000000,6003a04c) ret=6001e9e4 0806bb20:Ret user32.SetTimer() retval=00001003 ret=6001e9e4
in the trace, but never is the callback routine invoked. Without the above patch, the callback is invoked and does the painting.
2. A WM_KEYUP (wparam is VK_CNTL) message is generated and processed even though no key is touched on the keyboard. If "-debugmsg +key" is specified the message never appears, nor does any trace:key: entry.
Has anybody else seen or trapped these bugs?
Guy Albertelli <galberte@neo.lrun.com>