285974 - Crash when we use ATOK on the other of editor[@ nsWindow::OnIMEQueryCharPosition]

Status: RESOLVED FIXED

(Core :: Internationalization, defect)

Description

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

On other of editor(e.g., button), if we used ATOK17, mozilla crash.
This is regression of bug 278061.

Stack Signature	 nsWindow::OnIMEQueryCharPosition 2c93d345
Product ID	FirefoxTrunk
Build ID	2005031106
Trigger Time	2005-03-13 04:41:19.0
Platform	Win32
Operating System	Windows NT 5.1 build 2600
Module	FIREFOX.EXE + (00108118)
URL visited	
User Comments	
Since Last Crash	13136 sec
Total Uptime	32478 sec
Trigger Reason	Access violation
Source File, Line No.
c:/builds/tinderbox/Fx-Trunk/WINNT_5.0_Depend/mozilla/widget/src/windows/nsWindow.cpp,
line 6740
Stack Trace 	
nsWindow::OnIMEQueryCharPosition 
[c:/builds/tinderbox/Fx-Trunk/WINNT_5.0_Depend/mozilla/widget/src/windows/nsWindow.cpp,
line 6740]
nsWindow::OnIMERequest 
[c:/builds/tinderbox/Fx-Trunk/WINNT_5.0_Depend/mozilla/widget/src/windows/nsWindow.cpp,
line 6577]
nsWindow::ProcessMessage 
[c:/builds/tinderbox/Fx-Trunk/WINNT_5.0_Depend/mozilla/widget/src/windows/nsWindow.cpp,
line 4795]
nsWindow::WindowProc 
[c:/builds/tinderbox/Fx-Trunk/WINNT_5.0_Depend/mozilla/widget/src/windows/nsWindow.cpp,
line 1419]
USER32.dll + 0x8709 (0x77cf8709)
USER32.dll + 0x87eb (0x77cf87eb)
USER32.dll + 0xb368 (0x77cfb368)
USER32.dll + 0xb3b4 (0x77cfb3b4)
ntdll.dll + 0xeae3 (0x7c94eae3)
USER32.dll + 0xff42 (0x77cfff42)
ATOK17W.IME + 0x133dde (0x5ed53dde)
0x0033002e

6727   if ((charPosition != 0 && charPosition != sIMECursorPosition) ||
6728       charPosition > IME_MAX_CHAR_POS) {
6729     *oResult = FALSE;
6730     return PR_FALSE;
6731   }
6732   POINT pt;
6733   pt.x = sIMECompCharPos[charPosition].left;
6734   pt.y = sIMECompCharPos[charPosition].top;
6735   ::ClientToScreen(mWnd, &pt);
6736   pCharPosition->pt = pt;
6737 
6738   pCharPosition->cLineHeight = sIMECaretHeight;
6739 
6740   ::GetWindowRect(mWnd, &pCharPosition->rcDocument);
6741 
6742   *oResult = TRUE;
6743   return PR_TRUE;
6744 }

I will create the patch for this bug. Please wait.

Comment 1

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Patch rv1.0

The patch fixes the crash and stops to return invalid value when the active
widget is not editor and the composition string is empty.

Comment 2

[timeless]

Comment on attachment 177373 [details] [diff] [review]
Patch rv1.0

i don't like mGotValue, please explain what it is so that a native speaker can
try to come up with a better variable name.

+  // If the char positions are not cached, we should not return.
not return? do you mean return failure? because it looks like you do return:

+  // Because in this case, the active widget is not editor.
+  if (!sIMECompCharPos) {
+    *oResult = FALSE;
+    return PR_FALSE;
+  }

Comment 3

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Robert:

Please help me for the variable name.
See comment 2.

In Reply to comment 2
> not return? do you mean return failure? because it looks like you do return:
The values not return by *aData when the *oResult is FALSE.
Do you have an idea for better comment?

Comment 4

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I assume it means "PR_TRUE if the current caret rect was successfully stored in
mCaretRect, otherwise PR_FALSE". Maybe just call it mRectIsValid?

Comment 5

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Yes. You're right. Thank you!

Comment 6

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Patch rv1.1

Comment 7

[timeless]

h

Comment 8

[timeless]

Comment on attachment 177716 [details] [diff] [review]
Patch rv1.1

mozilla/widget/public/nsGUIEvent.h	3.117
mozilla/editor/libeditor/base/nsEditor.cpp	1.439
mozilla/widget/src/windows/nsWindow.cpp 	3.546

Comment 9

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Thank you, timeless and robert!

-> FIXED