http://bugs.winehq.org/show_bug.cgi?id=15318
Dylan Smith dylan.ah.smith@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dylan.ah.smith@gmail.com
--- Comment #7 from Dylan Smith dylan.ah.smith@gmail.com 2008-10-21 01:44:56 --- I used AppDb's HOWTO for installing and running Chrome.
I found that the workaround for this issue is: $sh winetricks riched30
The following just caused chrome to crash: $sh winetricks riched20
The problem seems to be that there are 2 WM_PASTE messages being sent: #1 is sent on WM_KEYDOWN probably by chrome's window proc for the richedit control #2 is sent on WM_CHAR by Wine's window proc
It seems like chrome replaced the window proc for the richedit control and falls back on Wine's for some functionality.
Commenting out the SendMessageW(editor->hWnd, WM_PASTE, 0, 0) in the WM_CHAR case for richedit's window proc does fix the problem for chrome, which isn't a fix, but confirms some of my analysis.
I suspect that pasting should be done in WM_KEYDOWN, and that is why chrome avoids calling Wine's window proc for this message when it is for a Ctrl-v.
I still need to make a test to fix this properly, so I will need to figure out how to simulate a Ctrl-V with a keyboard message through SendMessage. The richedit tests have a comment that says "Native riched20 won't obey WM_CHAR messages or WM_KEYDOWN/WM_KEYUP messages, probably because it inspects the keyboard state itself."