Skip to contentSkip to conversation

qt5multimedia: fix compile without opengl

PATCH New Peter Seiderer 2 participants 4 messages Jan 28, 2015 – Feb 3, 2015

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/2 qt5multimedia: fix compile without opengl

Hello Thomas,

Quoted 39 lines · Thomas Petazzoni

Dear Peter Seiderer,

Quoted 35 lines · Peter Seiderer
package/qt5/qt5multimedia/0001-qpaintervideosurface-fix-compile-without-opengl.patch new file +29 -0
new file mode 100644
index 0000000..578b79e
@@ -0,0 +1,39 @@
+From bca9c8786da5b5bca47b873720cb0d576219d4a9 Mon Sep 17 00:00:00 2001
+From: Peter Seiderer <[email protected]>
+Date: Fri, 23 Jan 2015 18:58:29 +0100
+Subject: [PATCH] qpaintervideosurface: fix compile without opengl
+
+Signed-off-by: Peter Seiderer <[email protected]>
+---
+ src/multimediawidgets/qpaintervideosurface.cpp | 4 ++++
+ 1 file changed, 4 insertions(+)
+
+diff --git a/src/multimediawidgets/qpaintervideosurface.cpp b/src/multimediawidgets/qpaintervideosurface.cpp
+index 3a880de..667ecd3 100644
+--- a/src/multimediawidgets/qpaintervideosurface.cpp
++++ b/src/multimediawidgets/qpaintervideosurface.cpp
+@@ -95,8 +95,10 @@ QVideoSurfaceGenericPainter::QVideoSurfaceGenericPainter()
+         << QVideoFrame::Format_RGB32
+         << QVideoFrame::Format_ARGB32
+         << QVideoFrame::Format_RGB565;
++#if !defined(QT_NO_OPENGL) && !defined(QT_OPENGL_ES_1_CL) && !defined(QT_OPENGL_ES_1)
+     // The raster formats should be a subset of the GL formats.
+     if (QOpenGLContext::openGLModuleType() != QOpenGLContext::LibGLES)
++#endif
+         m_imagePixelFormats << QVideoFrame::Format_RGB24;
+ }
+ 
+@@ -137,8 +139,10 @@ QAbstractVideoSurface::Error QVideoSurfaceGenericPainter::start(const QVideoSurf
+     const QAbstractVideoBuffer::HandleType t = format.handleType();
+     if (t == QAbstractVideoBuffer::NoHandle) {
+         bool ok = m_imageFormat != QImage::Format_Invalid && !m_imageSize.isEmpty();
Quoted 8 lines

package/qt5/qt5multimedia/0001-qpaintervideosurface-fix-compile-without-opengl.patch line 33

Quoted 4 lines
package/qt5/qt5multimedia/0001-qpaintervideosurface-fix-compile-without-opengl.patch new file +4 -0
++#if !defined(QT_NO_OPENGL) && !defined(QT_OPENGL_ES_1_CL) && !defined(QT_OPENGL_ES_1)
+         if (QOpenGLContext::openGLModuleType() == QOpenGLContext::LibGLES)
+             ok &= format.pixelFormat() != QVideoFrame::Format_RGB24;
++#endif

This is a bit scarce on details, and the first hunk that just removes the if condition looks a bit suspicious. Could we instead get a patch accepted upstream?

The original code in the first hunk disables Format_RGB24 in case of QOpenGLContext::LibGLES, so it should be save to just remove the compare and so enable Format_RGB24 if OpenGL is disabled.

The original code in the second hunk disables Format_RGB24 in case of QOpenGLContext::LibGLES, so the right thing to do is to disable the if statement (checking on QOpenGLContext::LibGLES case) and the next line which sets ok to false in case of Format_RGB24...

But I searched upstream git, there is already a fix in 5.4.1/dev branch fixing this problem ([1]), but as I think doing it the wrong way, disabling Format_RGB24 for the non-OpenGL case ([2]), but a follow up patch fixing the issue is on its way upstream ([3])...

Regards,
Peter

[1] https://qt.gitorious.org/qt/qtmultimedia/commit/2b181d546970d18a48a0f36f5d1a22418b61cd4d
[2] https://codereview.qt-project.org/101817
[3] https://codereview.qt-project.org/104811

Quoted 8 lines

Thanks,

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#1 lore raw

PATCH v2 1/2 qt5multimedia: fix compile without opengl

Quoted 4 lines · Peter Seiderer

But I searched upstream git, there is already a fix in 5.4.1/dev branch fixing this problem ([1]), but as I think doing it the wrong way, disabling Format_RGB24 for the non-OpenGL case ([2]), but a follow up patch fixing the issue is on its way upstream ([3])...

Could you then submit those two upstream patches? Either separately or
squashed, but with a reference to the upstream commit and your SoB.

Regards,
Arnout

Quoted 22 lines

Regards,
Peter

[1] https://qt.gitorious.org/qt/qtmultimedia/commit/2b181d546970d18a48a0f36f5d1a22418b61cd4d
[2] https://codereview.qt-project.org/101817
[3] https://codereview.qt-project.org/104811

Quoted 8 lines

Thanks,

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
#2 lore raw

PATCH v2 1/2 qt5multimedia: fix compile without opengl

Hello Arnout,

Quoted 6 lines

Gesendet: Dienstag, 03. Februar 2015 um 17:21 Uhr
Von: "Arnout Vandecappelle" <[email protected]>
An: "Peter Seiderer" <[email protected]>, "Thomas Petazzoni" <[email protected]>
Cc: buildroot at busybox.net
Betreff: Re: [Buildroot] [PATCH v2 1/2] qt5multimedia: fix compile without opengl

Quoted 9 lines
Quoted 4 lines · Peter Seiderer

But I searched upstream git, there is already a fix in 5.4.1/dev branch fixing this problem ([1]), but as I think doing it the wrong way, disabling Format_RGB24 for the non-OpenGL case ([2]), but a follow up patch fixing the issue is on its way upstream ([3])...

Could you then submit those two upstream patches? Either separately or
squashed, but with a reference to the upstream commit and your SoB.

Mhh, I understand the reasoning for grabbing upstream patches, but in this case I am not sure its worth the work to possibly rebase the upstream version (in case they do not apply cleanly, or to squash them), to get nearly the same result as with the suggested patch, the fix for [2] with [3] is derived work of my suggested buildroot patch ;-)

Less work would be to just add the upstream references to the commit message and/or the patch...

Any preferred solution?

Regards,
Peter

Quoted 35 lines

Regards,
Arnout

Quoted 22 lines

Regards,
Peter

[1] https://qt.gitorious.org/qt/qtmultimedia/commit/2b181d546970d18a48a0f36f5d1a22418b61cd4d
[2] https://codereview.qt-project.org/101817
[3] https://codereview.qt-project.org/104811

Quoted 8 lines

Thanks,

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
#3 lore raw

PATCH v2 1/2 qt5multimedia: fix compile without opengl

Quoted 20 lines · Peter Seiderer

Hello Arnout,

Quoted 3 lines

Gesendet: Dienstag, 03. Februar 2015 um 17:21 Uhr
Von: "Arnout Vandecappelle" <[email protected]>
An: "Peter Seiderer" <[email protected]>, "Thomas Petazzoni"

<[email protected]>

Quoted 2 lines

Cc: buildroot at busybox.net
Betreff: Re: [Buildroot] [PATCH v2 1/2] qt5multimedia: fix compile without

opengl

Quoted 10 lines · Peter Seiderer
Quoted 4 lines

But I searched upstream git, there is already a fix in 5.4.1/dev branch fixing this problem ([1]), but as I think doing it the wrong way, disabling Format_RGB24 for the non-OpenGL case ([2]), but a follow up patch fixing the issue is on its way upstream ([3])...

Could you then submit those two upstream patches? Either separately or
squashed, but with a reference to the upstream commit and your SoB.

Quoted 8 lines

Mhh, I understand the reasoning for grabbing upstream patches, but in this case I am not sure its worth the work to possibly rebase the upstream version (in case they do not apply cleanly, or to squash them), to get nearly the same result as with the suggested patch, the fix for [2] with [3] is derived work of my suggested buildroot patch ;-)

Less work would be to just add the upstream references to the commit message and/or the patch...

Ah OK, to me it looked as if the upstream patches were handling the situation
differently, but I hadn't taken the time to look at them in detail.

If you're doing basically the same thing, then your patch is OK as if of
course. The upstream references are still useful but it could be applied without it.

So basically:

Signature · 6 lines
Arnout Vandecappelle arnout at mind be
Senior Embedded Software Architect +32-16-286500
Essensium/Mind http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint: 7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F