183768 - Crash in nsWindow if mouse clicked before parentWidget is available

Status: RESOLVED FIXED

(Core :: Layout, defect, P1)

Description

[Angelo Nunes]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.0.1) Gecko/20020823 Netscape/7.0
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.0.1) Gecko/20020823 Netscape/7.0

 HWND parentWnd = ::GetParent(mWnd);
    if (parentWnd) {
      nsWindow* parentWidget = GetNSWindowPtr(parentWnd);
      if (parentWidget->mIMEIsComposing && nsWindow::uWM_MSIME_MOUSE) {

parentWnd is valid but parentWidget is NULL... crashola...
Happens if you click on the gecko view window too soon.

Reproducible: Always

Steps to Reproduce:
1.Launch a gecko viewer
2.click on the window immediately after its viewable
3.

Actual Results:  
Crash

Expected Results:  
return

Comment 1

[rpotts (gone)]

hey rod

can you take a look at this one?

thanks,
-- rick

Comment 2

[Angelo Nunes]

Patch I applied to my branch:

Index: nsWindow.cpp
===================================================================
RCS file: /cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v
retrieving revision 3.453
diff -u -r3.453 nsWindow.cpp
--- nsWindow.cpp        25 Oct 2002 00:03:47 -0000      3.453
+++ nsWindow.cpp        5 Dec 2002 19:48:20 -0000
@@ -6675,7 +6675,7 @@
     HWND parentWnd = ::GetParent(mWnd);
     if (parentWnd) {
       nsWindow* parentWidget = GetNSWindowPtr(parentWnd);
-      if (parentWidget->mIMEIsComposing && nsWindow::uWM_MSIME_MOUSE) {
+      if (parentWidget && parentWidget->mIMEIsComposing && nsWindow::uWM_MSIME_
MOUSE) {
         if (parentWidget->IMECompositionHitTest(aEventType, &ptPos))
           if (parentWidget->HandleMouseActionOfIME(aAction, &ptPos))
             return PR_TRUE;

Comment 3

[rods (gone)]

Attachment: patch

Simple fix, just make sure we have a valid pointer

Comment 4

[Johnny Stenback (:jst)]

Comment on attachment 108395 [details] [diff] [review]
patch

sr=jst

Comment 5

[dcone (gone)]

Comment on attachment 108395 [details] [diff] [review]
patch

r=dcone

Comment 6

[rods (gone)]

Requesting for inclusion in 1.3a, zero risk, it is an extra valid pointer check.

Comment 7

[Asa Dotzler [:asa]]

Comment on attachment 108395 [details] [diff] [review]
patch

I assume the "blocking1.3a" was intended as a request to land this patch (which
should be done with the "approval1.3a" flag in the patch manager). Removed
blocking request and a=asa for checkin to 1.3a.

Comment 8

[rods (gone)]

fixed