297683 - WINCE fix up

Status: RESOLVED FIXED

(Firefox Build System :: General, defect)

Description

[Doug Turner (:dougt)]

Comment 1

[Doug Turner (:dougt)]

Attachment: patch v.1

* Fixes some build problems when trying to build non-static.  (removing link
libraries that don't exist on Windows CE)

* Taking xxx advice and making GetNAME always return a GDI forcing us to use
GetTextMetrics instead of GetFontData (which doesn't exist on Windows CE)

* Windows CE on ARM doesn't support C++ exceptions.  In the windows plugin code
uses catches to ensure that the plugins don't crash mozilla.  On windows ce on
arm, we have no choice but to disable CALL_SAFETY_ON.

* On Windows CE, VerQueryValue internal structure can not be modified so
callers must do a bit of string conversion.

* Disabling more code in nsOSHelperAppService for Windows CE.

* Enabling the Windows Registry code in XPCOM since the Window CE shunt now
supports the limited set needed (all but notification).

*

Comment 2

[Doug Turner (:dougt)]

Comment on attachment 186224 [details] [diff] [review]
patch v.1

in the comment above, the mystery man "xxx" is Roger B. Sidje.

Comment 3

[Daniel Veditz [:dveditz]]

Comment on attachment 186224 [details] [diff] [review]
patch v.1

sr=dveditz
I'd prefer if nsWindowsRegKey::StartWatching() put the short WINCE ifdef at the
top and got it out of the way like the other places.

Comment 4

[Doug Turner (:dougt)]

np, will do.  thx

Comment 5

[Ere Maijala (slow)]

Comment on attachment 186224 [details] [diff] [review]
patch v.1

"Me too" on the positioning. r=emaijala

Comment 6

[Doug Turner (:dougt)]

Checking in gfx/src/windows/Makefile.in;
/cvsroot/mozilla/gfx/src/windows/Makefile.in,v  <--  Makefile.in
new revision: 1.27; previous revision: 1.26
done
Checking in gfx/src/windows/nsFontMetricsWin.cpp;
/cvsroot/mozilla/gfx/src/windows/nsFontMetricsWin.cpp,v  <--  nsFontMetricsWin.cpp
new revision: 3.229; previous revision: 3.228
done
Checking in modules/plugin/base/src/nsPluginSafety.h;
/cvsroot/mozilla/modules/plugin/base/src/nsPluginSafety.h,v  <--  nsPluginSafety.h
new revision: 1.14; previous revision: 1.13
done
Checking in modules/plugin/base/src/nsPluginsDirWin.cpp;
/cvsroot/mozilla/modules/plugin/base/src/nsPluginsDirWin.cpp,v  <-- 
nsPluginsDirWin.cpp
new revision: 1.35; previous revision: 1.34
done
Checking in modules/plugin/samples/default/windows/Makefile.in;
/cvsroot/mozilla/modules/plugin/samples/default/windows/Makefile.in,v  <-- 
Makefile.in
new revision: 1.16; previous revision: 1.15
done
Checking in uriloader/exthandler/win/nsOSHelperAppService.cpp;
/cvsroot/mozilla/uriloader/exthandler/win/nsOSHelperAppService.cpp,v  <-- 
nsOSHelperAppService.cpp
new revision: 1.63; previous revision: 1.62
done
Checking in xpcom/ds/Makefile.in;
/cvsroot/mozilla/xpcom/ds/Makefile.in,v  <--  Makefile.in
new revision: 1.98; previous revision: 1.97
done
Checking in xpcom/ds/nsWindowsRegKey.cpp;
/cvsroot/mozilla/xpcom/ds/nsWindowsRegKey.cpp,v  <--  nsWindowsRegKey.cpp
new revision: 1.4; previous revision: 1.3
done

Comment 7

[Doug Turner (:dougt)]

I am going to reopen this bug for further changes related to the WinCE port.

Comment 8

[Doug Turner (:dougt)]

Attachment: patch v.2

This patch does two things:

1) Correctly handles dropdown menus.  In earlier version of nsWindow for WinCE
we basically totally ignored handling events that were outside of the popup. 
The result was that if you happen to popup a menu, it wouldn't go away unless
you clicked on a item.	If there we no ideas, well you were hosed.

2) Correctly handles the WM_WINDOWPOSCHANGED event.  Prior to this, if the
browser ever was resized or moved, we would ignore the event.  We can not
process the WINDOWPLACEMENT and instead just always focus.

Comment 9

[Doug Turner (:dougt)]

Comment on attachment 187438 [details] [diff] [review]
patch v.2

Please ignore the Makefile.in change.

Comment 10

[Daniel Veditz [:dveditz]]

Comment on attachment 187438 [details] [diff] [review]
patch v.2

sr=dveditz

Comment 11

[Ere Maijala (slow)]

Comment on attachment 187438 [details] [diff] [review]
patch v.2

r=emaijala

Comment 12

[Doug Turner (:dougt)]

Checking in src/windows/nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.561; previous revision: 3.560
done
Checking in src/windows/nsWindow.h;
/cvsroot/mozilla/widget/src/windows/nsWindow.h,v  <--  nsWindow.h
new revision: 3.202; previous revision: 3.201
done

Leaving open for further fixups.  

Comment 13

[Doug Turner (:dougt)]

Attachment: patch v.3

Implements DataToBitmap.  We do not handle the cursor translucency, but it is
more then what we do currently.  Basically we are moving around ifdef's here.

Fixes GetSysFontInfo.  For windows CE we correctly return the System Font.

Comment 14

[Ere Maijala (slow)]

Comment on attachment 188814 [details] [diff] [review]
patch v.3

r=emaijala

Comment 15

[Daniel Veditz [:dveditz]]

Comment on attachment 188814 [details] [diff] [review]
patch v.3

sr=dveditz

Comment 16

[Doug Turner (:dougt)]

Checking in gfx/src/windows/nsDeviceContextWin.cpp;                            
      /cvsroot/mozilla/gfx/src/windows/nsDeviceContextWin.cpp,v  <-- 
nsDeviceContextWin.cppnew revision: 3.121; previous revision: 3.120            
                            done                                               
                                  Checking in widget/src/windows/nsWindow.cpp; 
                                       
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp          
      new revision: 3.564; previous revision: 3.563                            
            done                                                               
                  

thanks.  leaving open for further changes for wince fix up.

Comment 17

[Doug Turner (:dougt)]

Attachment: patch v.4

This patch does the follow:

*)  First it allows hardware key that various devices have (so called HOTKEYs)
be mapped into the DOM (as F-KEYS).  This will allow application developers use
the standard hardware buttons on devcies in some interesting way.  This is
controlled via a preference.  it is to be defaulted on -- we will override the
hardware buttons and pass these events on.

*) Secondly, i have added a preference which controls if the top level windows
have a (X) box or a (OK) box in their upper right hand corner.	The difference
between these is that an (X) box only puts the top level window in the
background whereas the (OK) box actually is suppose close and destory the
window.  It is very important for developers (a time save) to have the ablity
to directly kill the window.  Without this, you have to go to Settings / System
/ Memory / Running Programs, then find your app and press "Stop".  

This patch is WINCE only -- everything is within the correct #defines.

Comment 18

[Daniel Veditz [:dveditz]]

Comment on attachment 189851 [details] [diff] [review]
patch v.4

>+#include "aygshell.h"

Something you try not to walk on ("eggshell")?

> gProcUndergisterFunc

Undergister?

Do we really want to rely on undocumented functions? We see what happens to our
own embeddors when they do that...

>@@ -1558,20 +1630,23 @@ nsWindow::StandardWindowCreate(nsIWidget
>+#ifdef WINCE
>+    if (isMozWindowTakingFocus && aEventType == NS_GOTFOCUS)
>+      MapHardwareButtons(mWnd);
>+#endif

If you have to do this, do we need to un-map the hardward buttons if we lose
focus?

sr=dveditz

Comment 19

[Doug Turner (:dougt)]

Comment on attachment 189851 [details] [diff] [review]
patch v.4

I will fix up the typo in the var name.

It is undocument, but that is what you have to do if you want to override hw
buttons.

also, these keys are unmapped if we loose focus -- so if another app comes in,
the hw keys work just fine.

Comment 20

[Ere Maijala (slow)]

Comment on attachment 189851 [details] [diff] [review]
patch v.4

r=emaijala

Comment 21

[Doug Turner (:dougt)]

Checking in src/windows/nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.565; previous revision: 3.564
done

Comment 22

[Doug Turner (:dougt)]

Attachment: patch v.5

really simply wince ifdefs.

Comment 23

[Daniel Veditz [:dveditz]]

Comment on attachment 191988 [details] [diff] [review]
patch v.5

r/sr=dveditz

Comment 24

[Doug Turner (:dougt)]

Checking in nsFontMetricsWin.cpp;
/cvsroot/mozilla/gfx/src/windows/nsFontMetricsWin.cpp,v  <--  nsFontMetricsWin.cpp
new revision: 3.232; previous revision: 3.231
done
Checking in nsGfxFactoryWin.cpp;
/cvsroot/mozilla/gfx/src/windows/nsGfxFactoryWin.cpp,v  <--  nsGfxFactoryWin.cpp
new revision: 3.40; previous revision: 3.39
done

Comment 25

[Doug Turner (:dougt)]

Attachment: patch v.6

This patch does two things:
a) enables clipboard support on Windows CE.  It was pretty easy to port.  we
just don't support file clipboard type nor do we know what to do with drag and
drop.  Hence there are a few ifdef's sprinkled through out.  Also, i noticed
that the WinCE compiler chokeds on ::Global* for some reason.  Since these do
not need to be namespaced, I removed the "::".

b) detects windows ce tap-and-hold events via SHRecognizeGesture.  When we
discover that this occurs we fire a right mouse click.

Comment 26

[Daniel Veditz [:dveditz]]

Comment on attachment 192305 [details] [diff] [review]
patch v.6

r/sr=dveditz

Comment 27

[Doug Turner (:dougt)]

Checking in src/build/nsWinWidgetFactory.cpp;
/cvsroot/mozilla/widget/src/build/nsWinWidgetFactory.cpp,v  <-- 
nsWinWidgetFactory.cpp
new revision: 1.66; previous revision: 1.65
done
Checking in src/windows/Makefile.in;
/cvsroot/mozilla/widget/src/windows/Makefile.in,v  <--  Makefile.in
new revision: 3.26; previous revision: 3.25
done
Checking in src/windows/nsClipboard.cpp;
/cvsroot/mozilla/widget/src/windows/nsClipboard.cpp,v  <--  nsClipboard.cpp
new revision: 1.95; previous revision: 1.94
done
Checking in src/windows/nsDataObj.cpp;
/cvsroot/mozilla/widget/src/windows/nsDataObj.cpp,v  <--  nsDataObj.cpp
new revision: 1.69; previous revision: 1.68
done
Checking in src/windows/nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.568; previous revision: 3.567
done

Comment 28

[Doug Turner (:dougt)]

Attachment: patch v.7

1) removes unused variable mBorderlessParent.
2) fixes a few bugs whereby I #ifdef out a case in a switch statement resulting
the default action being taken.
3) allowing gesture animation via a flag to SHRecognizeGesture.
4) returning a positive value to the WM_ACTIVATE message as the caller excepts
that.
5) GetWindowPlacement() is not implemented in the wince shunt library.	So, i
could remove some of the #ifndef WINCE code when handing WM_WINDOWPOSCHANGED.
6) Adjusting the window position when WM_SETTINGCHANGE is called with
SPI_SETWORKAREA.  This is a windows ce thing. Basically, you can go to full
screen and a WM_SETTINGCHANGE message is fired.
7) Lastly, when dealing with popups, we want to ignore the activate message
when deciding to roll up the popup.

Comment 29

[Doug Turner (:dougt)]

Checking in src/windows/nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.572; previous revision: 3.571
done
Checking in src/windows/nsWindow.h;
/cvsroot/mozilla/widget/src/windows/nsWindow.h,v  <--  nsWindow.h
new revision: 3.206; previous revision: 3.205
done

Comment 30

[Doug Turner (:dougt)]

Attachment: patch v.8

This patch does the following:
  * Removes #ifndef WINCE around the OleRegisterMgr.  I have implemented the
required OLE functions in the windows ce shunt layer.

  * In nsWindow::InitEvent, i moved code from outside of the if-statement as it
was only being used within the true block.

  * In nsWindow::EventIsInsideWindow, we need to treat Activation events are
non-fatal.  That is to say, if a popup window gets an activation event, it
shouldn't cause the popup to go away on windows ce.

  * Removing the #ifndef WINCE around the dialog code.	Now eWindowType_dialog
really will get you a dialog just like on Windows.

  * I have stubbed out MA_* defines in the WinCE shunt layer.  This allowed us
to remove the #ifdef in nsWindow::ProcessMessage.

Comment 31

[Daniel Veditz [:dveditz]]

Comment on attachment 194418 [details] [diff] [review]
patch v.8

r/sr=dveditz

Comment 32

[Doug Turner (:dougt)]

Checking in nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.580; previous revision: 3.579
done

thanks.

Comment 33

[Doug Turner (:dougt)]

Attachment: patch v.9

This patch address the follow:
1) For windows CE, we do not ever want to ask for a font with ANSI_CHARSET.  We
should try to always use the default charset.
2) Also, CLIP_TURNOFF_FONTASSOCIATION (a #define in this file), is not support
on WINCE.
3) Since WINCE has an all unicode backend, we can simple never use the
nsFontWinNonUnicode class.  I have only ifdef'ed the instantiation of these
objects, in a later patch, I hope to ifdef out the nsFontWinNonUnicode all
together.
4) SubstituteChars need to not sub anything.  If the glyph isn't present, then
it is a font problem and a [] will apper.
5) Lastly, I included a patch to remove the #ifdef's from nsOSHelperAppService.
 The shunt library now supports all the functionality that is needed.

Comment 34

[Doug Turner (:dougt)]

Comment on attachment 195163 [details] [diff] [review]
patch v.9

I will get rid of the tabs in the patch... Sorry.

Comment 35

[Daniel Veditz [:dveditz]]

Comment on attachment 195163 [details] [diff] [review]
patch v.9

r/sr=dveditz

Comment 36

[Doug Turner (:dougt)]

Attachment: patch v.a

Patch does the following:
1) Enables native theme support on WinCE so that things look more clean --
specifically scrollbars looks better.
2) SetGraphicsMode isn't implemented on WinCE.	I ifdef'ed it out for WINCE.  I
think stuart, vlad, and I will have to really determine if i need to implement
this in the WinCE shunt.  For now, keep everything building.
3) I enabled IMM support on WINCE, although it is very rarely used.  This
allows us to remove some ifdef's.
4) I am remapping the hardward keys to new values.
5) I am also allowing the SmartPhone softkey's and the Back button to be
remapped.

Comment 37

[Daniel Veditz [:dveditz]]

Comment on attachment 199393 [details] [diff] [review]
patch v.a

r/sr=dveditz

Comment 38

[Doug Turner (:dougt)]

Checking in widget/src/build/nsWinWidgetFactory.cpp;
/cvsroot/mozilla/widget/src/build/nsWinWidgetFactory.cpp,v  <-- 
nsWinWidgetFactory.cpp
new revision: 1.68; previous revision: 1.67
done
Checking in widget/src/windows/Makefile.in;
/cvsroot/mozilla/widget/src/windows/Makefile.in,v  <--  Makefile.in
new revision: 3.29; previous revision: 3.28
done
Checking in widget/src/windows/nsNativeThemeWin.cpp;
/cvsroot/mozilla/widget/src/windows/nsNativeThemeWin.cpp,v  <-- 
nsNativeThemeWin.cpp
new revision: 3.49; previous revision: 3.48
done
Checking in widget/src/windows/nsToolkit.cpp;
/cvsroot/mozilla/widget/src/windows/nsToolkit.cpp,v  <--  nsToolkit.cpp
new revision: 3.54; previous revision: 3.53
done
Checking in widget/src/windows/nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.583; previous revision: 3.582
done

Comment 39

[[e-mail]]

Comment on attachment 199393 [details] [diff] [review]
patch v.a

>@@ -2103,20 +2180,21 @@
>     toolkit->CallMethod(&info);
>     return NS_ERROR_FAILURE;
>   }
> 
>   if (mWnd) {
>     // Uniconify, if necessary
>     HWND toplevelWnd = GetTopLevelHWND(mWnd);
>     if (::IsIconic(toplevelWnd))
>       ::OpenIcon(toplevelWnd);
>     ::SetFocus(mWnd);
>+    MapHardwareButtons(mWnd);
>   }
>   return NS_OK;
> }
This one needs #ifdef WINCE too?

Comment 40

[Doug Turner (:dougt)]

Checking in nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.584; previous revision: 3.583
done

Comment 41

[Doug Turner (:dougt)]

Attachment: patch v.b

All Windows CE.

1) This only creates one softkey menu.  If we create more then one, this cause windows to be hidden between new window creations.  See bug 314552.

2) This also makes a improvement to mouse taps in a context menu.  Now we only test for tab-and-hold (a "Gesture") when we are not displaying a context menu.  If we don't do this, the check for the gesture makes the context menu go away.

3) We also hide the software input panel when we are creating a dialog or a window.  This insures that the windows ce SIP button never blocks the view of a XUL implemented SIP.

This is a -w diff since it is clearer what I am doing.

Comment 42

[Daniel Veditz [:dveditz]]

Comment on attachment 201632 [details] [diff] [review]
patch v.b

r=dveditz

Comment 43

[Doug Turner (:dougt)]

Checking in nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.589; previous revision: 3.588
done

ty.

Comment 44

[Doug Turner (:dougt)]

Attachment: patch v.c

1) enables the file picker

Comment 45

[Daniel Veditz [:dveditz]]

Comment on attachment 204850 [details] [diff] [review]
patch v.c

r/sr=dveditz

Comment 46

[Doug Turner (:dougt)]

Attachment: patch v.d

Patch:

a) allows remapping softkeys even if gProcUnregisterFunc does not exist -- which is the case on some devices.

b) removes gSoftKeyMenuBar as we can use mSoftKeyMenuBar.

c) fixes smartphone which require a real menu when dispatching events to the event loop.  SHCMBF_EMPTYBAR doesn't not work.  The menu is defined by the embedding application.

Comment 47

[Daniel Veditz [:dveditz]]

Comment on attachment 209764 [details] [diff] [review]
patch v.d

>+#ifdef WINCE
>+  // are we suppose to free mSoftKeyMenuBar.  MSDN isn't sure.
>+#endif

Please prefix this comment with XXX so it can be picked up in later scans.

"supposed" typo and s/\./?/

r=dveditz

Comment 48

[Doug Turner (:dougt)]

/cvsroot/mozilla/widget/src/windows/resource.h,v  <--  resource.h
new revision: 3.10; previous revision: 3.9
done
Checking in widget/src/windows/nsWindow.h;
/cvsroot/mozilla/widget/src/windows/nsWindow.h,v  <--  nsWindow.h
new revision: 3.215; previous revision: 3.214
done
Checking in widget/src/windows/nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.609; previous revision: 3.608
done


Thanks dan.

Comment 49

[Doug Turner (:dougt)]

Attachment: patch v.e

ensure that there is only one softkey menubar since having more results in a window minimization when an dialog is opened.

remove config.wince.useOKBtn pref.  it isn't used any longer.

Comment 50

[Daniel Veditz [:dveditz]]

Comment on attachment 217922 [details] [diff] [review]
patch v.e

>     if (!nsWindow::sIsPopupClassRegistered) {
>       // For older versions of Win32 (i.e., not XP), the registration will
>       // fail, so we have to re-register without the CS_XP_DROPSHADOW flag.
>-      wc.style = CS_DBLCLKS;
>       nsWindow::sIsPopupClassRegistered = nsToolkit::mRegisterClass(&wc);
>     }

This bit is not #ifdef WINCE, isn't this change going to break the thing
the comment says it's trying not to break? Or if it's ok now then should
the comment go away too?

Comment 51

[Doug Turner (:dougt)]

dan, youre right, for now, it should be #ifndef WINCE.

Comment 52

[Daniel Veditz [:dveditz]]

Comment on attachment 217922 [details] [diff] [review]
patch v.e

Discussed via IM, r=dveditz if the questioned bit is changed to:

>     if (!nsWindow::sIsPopupClassRegistered) {
 + #ifndef WINCE
>       // For older versions of Win32 (i.e., not XP), the registration will
>       // fail, so we have to re-register without the CS_XP_DROPSHADOW flag.
>       wc.style = CS_DBLCLKS;
 + #endif
>       nsWindow::sIsPopupClassRegistered = nsToolkit::mRegisterClass(&wc);
>     }

Comment 53

[Doug Turner (:dougt)]

Attachment: patch v.f

IME continues to make context menus and dropdowns not work as expected.  this patch ifdefs out msg handling of IME messages on WINCE.

This also reduces the differences between wince and win32 wrt to style and default window icon handling.

wince only

Comment 54

[Daniel Veditz [:dveditz]]

Comment on attachment 218852 [details] [diff] [review]
patch v.f

r=dveditz

Comment 55

[Doug Turner (:dougt)]

checked in. ty.

Comment 56

[Doug Turner (:dougt)]

Attachment: patch v.A1

Enables IME on windows ce.  this is required for any wince application to use IME.

Comment 57

[Doug Turner (:dougt)]

Comment on attachment 223799 [details] [diff] [review]
patch v.A1

that is the wrong patch.

Comment 58

[Doug Turner (:dougt)]

Attachment: patch v.A1 (really)

Comment 59

[Doug Turner (:dougt)]

Comment on attachment 223800 [details] [diff] [review]
patch v.A1 (really)

there is an extra } needed.  after: 
+        ImmEscapeW(NULL, hC, IME_ESC_SET_MODE, (LPVOID)IM_SPELL);

Comment 60

[Daniel Veditz [:dveditz]]

Comment on attachment 223800 [details] [diff] [review]
patch v.A1 (really)

sr=dveditz

Comment 61

[Doug Turner (:dougt)]

Attachment: patch v.A2

Windows ce only.

This fixes a regression caused my the last check-in (patch v.A1).  The regression caused WM_HOTKEY events from never occuring when a popup window (such as a xul menu) was displayed.  The reason for this is that the activation state on windows ce differs a bit from windows 32 (do to the lack of a cancelable mouse activation message).  Short story is that it seams we have to mask any activation events during popup handling.

Comment 62

[Daniel Veditz [:dveditz]]

Comment on attachment 232824 [details] [diff] [review]
patch v.A2

r=dveditz

Comment 63

[Doug Turner (:dougt)]

Comment on attachment 232824 [details] [diff] [review]
patch v.A2

wince only change.

Comment 64

[Mike Beltzner [:beltzner, not reading bugmail]]

Comment on attachment 232824 [details] [diff] [review]
patch v.A2

a=drivers, WinCE only change, totally cool by us.

Comment 65

[Doug Turner (:dougt)]

Checking in nsWindow.cpp;
/cvsroot/mozilla/widget/src/windows/nsWindow.cpp,v  <--  nsWindow.cpp
new revision: 3.569.2.33; previous revision: 3.569.2.32
done

Not fixed on the trunk.  Probably will need to be ported when windows ce works on the trunk again (it doesn't build right now).

Comment 66

[Doug Turner (:dougt)]

Attachment: patch v.A3

Windows CE ONLY

1. Use PostMessages as we do not care about the immediate result nor wish to wait.

2. Remove gSoftKeyMenuBar as we want all toplevel, dialog, and popup windows to have a softkey menu bar.

3. Reverting the SWP_NOACTIVATE flag to SetWindowPos.  It really seams that we can't click on the popup's without this.  Because of this change, I also reverting the check in DealWithPopups for the WM_ACTIVATE flag.

4. When handling the gesture in left mouse down, we always want to return zero.
Enabling WM_IME_ on windows ce.  The APIs are present, we just were ifdef'ing out the messaging handling.

Comment 67

[Daniel Veditz [:dveditz]]

Comment on attachment 233724 [details] [diff] [review]
patch v.A3

sr=dveditz

Comment 68

[Mike Connor [:mconnor]]

Comment on attachment 233724 [details] [diff] [review]
patch v.A3

a=mconnor on behalf of drivers for 1.8 branch checkin

Comment 69

[Doug Turner (:dougt)]

this was for 1.8 changes.   Marking fixed.