677712 - Make Qt port support No X11 build

Status: RESOLVED FIXED

(Core Graveyard :: Widget: Qt, defect)

Description

[Oleg Romashin (:romaxa)]

I think it would be nice to have Qt port compiling on NON-X11 platform like WebOS or something similar

Comment 1

[Oleg Romashin (:romaxa)]

Attachment: Compile Qt port on Non-X platform

Comment 2

[Oleg Romashin (:romaxa)]

Comment on attachment 551910 [details] [diff] [review]
Compile Qt port on Non-X platform

Wolfgang, not sure if dougt have time... could you also check this patch?

Comment 3

[Oleg Romashin (:romaxa)]

Attachment: Compile Qt port for Non-X platform

Removed duplicated define from nsObjectFrame, removed QtDbus (not in use)

Comment 4

[Oleg Romashin (:romaxa)]

Attachment: Fix Qt egl provider to work on nonX11 platform

Comment 5

[Oleg Romashin (:romaxa)]

Try build looks ok, with these patches:
http://tbpl.mozilla.org/?tree=Try&rev=2ba318c7a882

Comment 6

[Doug Turner (:dougt)]

Comment on attachment 553016 [details] [diff] [review]
Compile Qt port for Non-X platform


>+#elif defined(ANDROID) || defined(MOZ_WIDGET_QT)
> #  include "mozilla/plugins/NPEventAndroid.h"

really?

Comment 7

[Oleg Romashin (:romaxa)]

yep, definitely not NPEventX11.h...
Android is actually same as NPEventX11.h but does not include X stuff...
Another option is to rename NPEventX11.h -> NPEventUnix.h and add MOZ_X11 ifdefs inside...

Comment 8

[Doug Turner (:dougt)]

i think renaming might be better.  cc/ jst do you care one way or the other?

Comment 9

[Joe Drew (not getting mail)]

Comment on attachment 553038 [details] [diff] [review]
Fix Qt egl provider to work on nonX11 platform

Review of attachment 553038 [details] [diff] [review]:
-----------------------------------------------------------------

Looks fine as long as it passes try.

Comment 10

[Oleg Romashin (:romaxa)]

Attachment: Compile Qt port for Non-X platform

NPEventX11 renamed into NPEventUnix and ifdefed inside with MOZ_X11

Comment 11

[Doug Turner (:dougt)]

Comment on attachment 553821 [details] [diff] [review]
Compile Qt port for Non-X platform

Review of attachment 553821 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/plugins/base/nsNPAPIPlugin.cpp
@@ +2009,5 @@
>  
>    switch(variable) {
>  #if defined(XP_UNIX) && !defined(XP_MACOSX)
>    case NPNVxDisplay : {
> +#if defined(MOZ_WIDGET_GTK2) || (defined(MOZ_WIDGET_QT) && defined(MOZ_X11))

Shouldn't this just be:

#ifdef MOZ_X11 now?

::: widget/src/qt/nsToolkit.cpp
@@ +139,3 @@
>      return mSharedGC->handle();
> +#else
> +    return 0;

Is return zero bad?

http://mxr.mozilla.org/mozilla-central/source/widget/src/qt/nsWindow.cpp#803

Comment 12

[Oleg Romashin (:romaxa)]

Attachment: Compile Qt port for Non-X platform

Fixed MOZ_X11 define...
SharedGC returning 0 in not bad, also currently it is used only in ifdef WIN...
http://mxr.mozilla.org/mozilla-central/source/dom/plugins/base/nsPluginInstanceOwner.cpp#2939
http://mxr.mozilla.org/mozilla-central/source/dom/plugins/base/nsPluginInstanceOwner.cpp#2956

Comment 13

[Oleg Romashin (:romaxa)]

Attachment: Combined patch for push

Comment 14

[Oleg Romashin (:romaxa)]

Try build
http://tbpl.allizom.org/?tree=Try&usebuildbot=1&rev=a1998a0f883c

Comment 15

[Ed Morley [:emorley]]

Doesn't apply cleanly to inbound, due to bug 677920 landing. That bug changed the behaviour in that section, so not going to attempt to unbitrot, since I don't know if it changes what you need to do here.

Comment 16

[Oleg Romashin (:romaxa)]

Attachment: Updated to latest m-c tip

Comment 17

[Ed Morley [:emorley]]

Awesome, thanks :-)

http://hg.mozilla.org/integration/mozilla-inbound/rev/206ab0f10adf

Comment 18

[Matt Brubeck (:mbrubeck)]

http://hg.mozilla.org/mozilla-central/rev/206ab0f10adf

Comment 19

[Florian Hänel [:heeen]]

I am getting tons of these from this patch:

###!!! ASSERTION: Failed to make GL context current!: 'succeeded', file mozilla-central/gfx/thebes/GLContextProviderEGL.cpp, line 795

Comment 20

[Oleg Romashin (:romaxa)]

Sounds like mSurface check is not needed in Qt backend MakeCurrentImpl... will check it today quickly.

Comment 21

[Oleg Romashin (:romaxa)]

Ok, problem is in this line:
http://hg.mozilla.org/mozilla-central/rev/6e7449c449ba#l2.33
and mSurface check should not be performed for platform context.
I would suggest to file new bug.