Skip to contentSkip to conversation

qt5: bump to 5.1.1

PATCH Superseded Fatih Aşıcı 4 participants 8 messages Sep 18, 2013 – Sep 22, 2013

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/5 qt5: bump to 5.1.1

Sorry. I forgot to add mail list to cc.

Quoted 11 lines · On Wednesday 18 September 2013 20:04:37 you wrote:
Quoted 11 lines

-diff --git a/src/platformsupport/eglconvenience/eglconvenience.pri b/src/platformsupport/eglconvenience/eglconvenience.pri -index 188eb1c..8996cea 100644
---- a/src/platformsupport/eglconvenience/eglconvenience.pri -+++ b/src/platformsupport/eglconvenience/eglconvenience.pri -@@ -5,6 +5,7 @@ contains(QT_CONFIG,egl) {

(quoted diff) +0 -5
-     SOURCES += \
-         $$PWD/qeglconvenience.cpp \
-         $$PWD/qeglplatformcontext.cpp
-+    CONFIG += egl
-
Quoted 8 lines

Review

Quoted 5 lines
(quoted diff) +0 -5
-     contains(QT_CONFIG,xlib) {
-         HEADERS += \
---
-1.7.9.5
-

Why is this patch removed? Was this fixed in Qt upstream? I see you're adding a CONFIG += egl in another patch, but only for the X.org case.

It is fixed in upstream. The new patch is also sent to upstream and it is applied, too:

https://bugdiffs.qt-project.org/fisheye/changelog/Qt5-
qtbase?cs=ff7b57bbb936bc86a64aa2a947fa8984adf400c8

Quoted 15 lines
Quoted 15 lines
(quoted diff) +0 -0

b/package/qt5/qt5base/qt5base-mkspecs-files.patch index bfa9c15..e406cfc 100644

package/qt5/qt5base/qt5base-mkspecs-files.patch +1 -1
@@ -14,9 +14,9 @@ Index: b/mkspecs/devices/linux-buildroot-g++/qmake.conf
 
 ===================================================================
 --- /dev/null
 +++ b/mkspecs/devices/linux-buildroot-g++/qmake.conf
 
-@@ -0,0 +1,43 @@
+@@ -0,0 +1,44 @@
 
Quoted 6 lines

package/qt5/qt5base/qt5base-mkspecs-files.patch Index: b/mkspecs/devices/linux-buildroot-g++/qmake.conf

Quoted 4 lines
package/qt5/qt5base/qt5base-mkspecs-files.patch +1 -1
 +MAKEFILE_GENERATOR      = UNIX
 
-+CONFIG                  += incremental gdb_dwarf_index
++CONFIG                 += incremental gdb_dwarf_index

Is this change needed? Seems like white space change only, no?

I checked the mkspec files in old and new versions of upstream. Upstream's files have this change. So I changed our mkspec file to reduce diff output. I think, this will help checking again in new versions. Yes, it is only a white space change and not needed.

Quoted 35 lines
Quoted 35 lines
(quoted diff) +0 -0

b/package/qt5/qt5base/qt5base.mk index ddc5574..0106c2c 100644

package/qt5/qt5base/qt5base.mk +0 -0
@@ -84,13 +84,17 @@ QT5BASE_CONFIGURE_OPTS += $(if

$(BR2_PACKAGE_QT5BASE_DIRECTFB),-directfb,-no-dir

QT5BASE_DEPENDENCIES += $(if $(BR2_PACKAGE_QT5BASE_DIRECTFB),directfb)

ifeq ($(BR2_PACKAGE_QT5BASE_XCB),y)

package/qt5/qt5base/qt5base.mk +6 -2
-QT5BASE_CONFIGURE_OPTS += -xcb
+QT5BASE_CONFIGURE_OPTS += -xcb -system-xkbcommon
 
 QT5BASE_DEPENDENCIES   += \
 
 	libxcb \
 	xcb-util-wm \
 	xcb-util-image \
 	xcb-util-keysyms \
 
-	xlib_libX11
+	xlib_libX11 \
+	libxkbcommon
+ifeq ($(BR2_PACKAGE_QT5BASE_WIDGETS),y)
+QT5BASE_DEPENDENCIES   += xlib_libXext
+endif
 
 else
 QT5BASE_CONFIGURE_OPTS += -no-xcb
 endif
@@ -172,7 +176,7 @@ define QT5BASE_CONFIGURE_CMDS
 
Quoted 8 lines

package/qt5/qt5base/qt5base.mk define QT5BASE_CONFIGURE_CMDS · line 181

Quoted 6 lines
package/qt5/qt5base/qt5base.mk +1 -1
 		-sysroot $(STAGING_DIR) \
 		-plugindir /usr/lib/qt/plugins \
 		-no-rpath \
 
-		-nomake examples -nomake demos -nomake tests \
+		-nomake examples -nomake tests \

No more demos to disable?

"examples" and "demos" directories are merged in this new version. configure script fails with "-nomake demos".

Quoted 11 lines
Quoted 11 lines
package/qt5/qt5base/qt5base.mk +0 -0
 		-device buildroot \
package/qt5/qt5base/qt5base.mk +0 -0
 		-no-c++11 \
 		$(QT5BASE_CONFIGURE_OPTS) \
(quoted diff) +0 -0

b/package/qt5/qt5declarative/qt5declarative.mk index 603c0e7..943f33d 100644

package/qt5/qt5declarative/qt5declarative.mk +0 -0
@@ -39,6 +39,7 @@ ifeq ($(BR2_PREFER_STATIC_LIB),)
 
Quoted 4 lines

package/qt5/qt5declarative/qt5declarative.mk ifeq ($(BR2_PREFER_STATIC_LIB),) · line 43

Quoted 4 lines
package/qt5/qt5declarative/qt5declarative.mk +0 -0
 define QT5DECLARATIVE_INSTALL_TARGET_LIBS
 
 	cp -dpf $(STAGING_DIR)/usr/lib/libQt5Qml*.so.* $(TARGET_DIR)/usr/lib
 	cp -dpf $(STAGING_DIR)/usr/lib/libQt5Quick*.so.* 

$(TARGET_DIR)/usr/lib

Quoted 10 lines
Quoted 10 lines
package/qt5/qt5declarative/qt5declarative.mk +1 -0
+	cp -dpfr $(STAGING_DIR)/usr/lib/qt/plugins/accessible

$(TARGET_DIR)/usr/lib/qt/plugins

package/qt5/qt5declarative/qt5declarative.mk +0 -0
 	cp -dpfr $(STAGING_DIR)/usr/lib/qt/plugins/qml*
 	$(TARGET_DIR)/usr/lib/qt/plugins
 
 endef
 endif
Quoted 8 lines

package/qt5/qt5jsbackend/qt5jsbackend-fix-uclibc.patch

Quoted 6 lines
(quoted diff) +0 -0

b/package/qt5/qt5jsbackend/qt5jsbackend-fix-uclibc.patch deleted file mode 100644
index 50a6dbc..0000000

package/qt5/qt5jsbackend/qt5jsbackend-fix-uclibc.patch deleted +0 -0

So this patch was merged upstream? Or is no longer needed?

It is upstreamed, too.

Quoted 9 lines
Quoted 4 lines

ifeq ($(BR2_PACKAGE_QT5BASE_LICENSE_APPROVED),y)

-QT5QUICK1_CONFIGURE_OPTS += -opensource -confirm-license

Hum why does this gets removed from qt5quick1 but not the other packages? Maybe it's only needed in qt5base, but in this case, it should be removed from all other packages, and as a separate patch, since it's unrelated to the version bump, no?

OK. I will remove all of them with a separate patch.

Quoted 4 lines

Thanks a lot!

Thomas

Thanks,
Fatih

#1 lore raw

PATCH 1/5 qt5: bump to 5.1.1

Dear Fatih A??c?,

Quoted 23 lines · Fatih A??c?

Sorry. I forgot to add mail list to cc.

Quoted 11 lines · On Wednesday 18 September 2013 20:04:37 you wrote:
Quoted 11 lines

-diff --git a/src/platformsupport/eglconvenience/eglconvenience.pri b/src/platformsupport/eglconvenience/eglconvenience.pri -index 188eb1c..8996cea 100644
---- a/src/platformsupport/eglconvenience/eglconvenience.pri -+++ b/src/platformsupport/eglconvenience/eglconvenience.pri -@@ -5,6 +5,7 @@ contains(QT_CONFIG,egl) {

(quoted diff) +0 -5
-     SOURCES += \
-         $$PWD/qeglconvenience.cpp \
-         $$PWD/qeglplatformcontext.cpp
-+    CONFIG += egl
-
Quoted 8 lines

Review

Quoted 5 lines
(quoted diff) +0 -5
-     contains(QT_CONFIG,xlib) {
-         HEADERS += \
---
-1.7.9.5
-

Why is this patch removed? Was this fixed in Qt upstream? I see you're adding a CONFIG += egl in another patch, but only for the X.org case.

Quoted 5 lines

It is fixed in upstream. The new patch is also sent to upstream and it is applied, too:

https://bugdiffs.qt-project.org/fisheye/changelog/Qt5-
qtbase?cs=ff7b57bbb936bc86a64aa2a947fa8984adf400c8

Ok, thanks. This should have been part of the commit log :)

Quoted 21 lines
Quoted 14 lines
Quoted 14 lines
(quoted diff) +0 -0

b/package/qt5/qt5base/qt5base-mkspecs-files.patch index bfa9c15..e406cfc 100644

package/qt5/qt5base/qt5base-mkspecs-files.patch +1 -1
@@ -14,9 +14,9 @@ Index: b/mkspecs/devices/linux-buildroot-g++/qmake.conf
 
 ===================================================================
 --- /dev/null
 +++ b/mkspecs/devices/linux-buildroot-g++/qmake.conf
 
-@@ -0,0 +1,43 @@
+@@ -0,0 +1,44 @@
 
Quoted 6 lines

package/qt5/qt5base/qt5base-mkspecs-files.patch Index: b/mkspecs/devices/linux-buildroot-g++/qmake.conf

Quoted 4 lines
package/qt5/qt5base/qt5base-mkspecs-files.patch +1 -1
 +MAKEFILE_GENERATOR      = UNIX
 
-+CONFIG                  += incremental gdb_dwarf_index
++CONFIG                 += incremental gdb_dwarf_index

Is this change needed? Seems like white space change only, no?

Quoted 4 lines

I checked the mkspec files in old and new versions of upstream. Upstream's files have this change. So I changed our mkspec file to reduce diff output. I think, this will help checking again in new versions. Yes, it is only a white space change and not needed.

It's a bit early, so maybe my brain is not yet up and running, but if the upstream has changed this, and we don't care about whether there's a space or not, why does our patch contains something to change this?

Quoted 7 lines
Quoted 4 lines

package/qt5/qt5base/qt5base-mkspecs-files.patch

Quoted 2 lines
package/qt5/qt5base/qt5base-mkspecs-files.patch +1 -1
-		-nomake examples -nomake demos -nomake tests \
+		-nomake examples -nomake tests \

No more demos to disable?

"examples" and "demos" directories are merged in this new version. configure script fails with "-nomake demos".

Ok. I was of course suspecting that, merely pointing out that a more detailed commit log would have been nice.

Quoted 36 lines
Quoted 12 lines
Quoted 12 lines
package/qt5/qt5base/qt5base-mkspecs-files.patch +0 -0
 		-device buildroot \
 		-no-c++11 \
 		$(QT5BASE_CONFIGURE_OPTS) \
(quoted diff) +0 -0

b/package/qt5/qt5declarative/qt5declarative.mk index 603c0e7..943f33d 100644

package/qt5/qt5declarative/qt5declarative.mk +0 -0
@@ -39,6 +39,7 @@ ifeq ($(BR2_PREFER_STATIC_LIB),)
 
Quoted 4 lines

package/qt5/qt5declarative/qt5declarative.mk ifeq ($(BR2_PREFER_STATIC_LIB),) · line 43

Quoted 4 lines
package/qt5/qt5declarative/qt5declarative.mk +0 -0
 define QT5DECLARATIVE_INSTALL_TARGET_LIBS
 
 	cp -dpf $(STAGING_DIR)/usr/lib/libQt5Qml*.so.* $(TARGET_DIR)/usr/lib
 	cp -dpf $(STAGING_DIR)/usr/lib/libQt5Quick*.so.* 

$(TARGET_DIR)/usr/lib

Quoted 10 lines
Quoted 10 lines
package/qt5/qt5declarative/qt5declarative.mk +1 -0
+	cp -dpfr $(STAGING_DIR)/usr/lib/qt/plugins/accessible

$(TARGET_DIR)/usr/lib/qt/plugins

package/qt5/qt5declarative/qt5declarative.mk +0 -0
 	cp -dpfr $(STAGING_DIR)/usr/lib/qt/plugins/qml*
 	$(TARGET_DIR)/usr/lib/qt/plugins
 
 endef
 endif
Quoted 8 lines

package/qt5/qt5jsbackend/qt5jsbackend-fix-uclibc.patch

Quoted 6 lines
(quoted diff) +0 -0

b/package/qt5/qt5jsbackend/qt5jsbackend-fix-uclibc.patch deleted file mode 100644
index 50a6dbc..0000000

package/qt5/qt5jsbackend/qt5jsbackend-fix-uclibc.patch deleted +0 -0

So this patch was merged upstream? Or is no longer needed?

Quoted 1 line

It is upstreamed, too.

Ditto: to be mentioned in the commit log.

Quoted 10 lines
Quoted 8 lines
Quoted 3 lines

ifeq ($(BR2_PACKAGE_QT5BASE_LICENSE_APPROVED),y)

-QT5QUICK1_CONFIGURE_OPTS += -opensource -confirm-license

Hum why does this gets removed from qt5quick1 but not the other packages? Maybe it's only needed in qt5base, but in this case, it should be removed from all other packages, and as a separate patch, since it's unrelated to the version bump, no?

OK. I will remove all of them with a separate patch.

Thanks a lot! Definitely nice to see Qt5 moving.

Best regards,

Thomas

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
#2 lore raw

PATCH 1/5 qt5: bump to 5.1.1

Quoted 13 lines · Thomas Petazzoni
Quoted 11 lines
Quoted 9 lines
Quoted 7 lines

-@@ -0,0 +1,43 @@
+@@ -0,0 +1,44 @@

+MAKEFILE_GENERATOR = UNIX

-+CONFIG += incremental gdb_dwarf_index
++CONFIG += incremental gdb_dwarf_index

Is this change needed? Seems like white space change only, no?

I checked the mkspec files in old and new versions of upstream.

Upstream's

Quoted 1 line

files have this change. So I changed our mkspec file to reduce diff

Quoted 8 lines

output.

Quoted 1 line

I think, this will help checking again in new versions. Yes, it is

only

Quoted 1 line

a white space change and not needed.

It's a bit early, so maybe my brain is not yet up and running, but if the upstream has changed this, and we don't care about whether there's a space or not, why does our patch contains something to change this?

It is just a cosmetic change not to see the difference when comparing files :)

If the rule is not to touch anywhere unrelated to the version change, I can revert that part.

#3 lore raw

PATCH 1/5 qt5: bump to 5.1.1

Quoted 23 lines · Fatih A??c?
Quoted 13 lines · Thomas Petazzoni
Quoted 11 lines
Quoted 9 lines
Quoted 7 lines

-@@ -0,0 +1,43 @@
+@@ -0,0 +1,44 @@

+MAKEFILE_GENERATOR = UNIX

-+CONFIG += incremental gdb_dwarf_index
++CONFIG += incremental gdb_dwarf_index

Is this change needed? Seems like white space change only, no?

I checked the mkspec files in old and new versions of upstream.

Upstream's

Quoted 1 line

files have this change. So I changed our mkspec file to reduce diff

Quoted 8 lines

output.

Quoted 1 line

I think, this will help checking again in new versions. Yes, it is

only

Quoted 1 line

a white space change and not needed.

It's a bit early, so maybe my brain is not yet up and running, but if the upstream has changed this, and we don't care about whether there's a space or not, why does our patch contains something to change this?

Quoted 5 lines

It is just a cosmetic change not to see the difference when comparing files :)

If the rule is not to touch anywhere unrelated to the version change, I can revert that part.

I think it's a good idea to keep it. The closer that our mkspecs stays
to one of the upstream examples, the better.

Perhaps we could try upstreaming our mkspecs? Does that make sense?

Regards,
Arnout

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
#4 lore raw

PATCH 1/5 qt5: bump to 5.1.1

Dear Arnout Vandecappelle,

Quoted 1 line · Arnout Vandecappelle

Perhaps we could try upstreaming our mkspecs? Does that make sense?

Our mkspecs is really meant to be a template that has to be tuned by the qt5base.mk makefile, and therefore outside Buildroot itself I believe its meaning is not so important. We could even carry it within package/qt5/qt5base/ rather than as a patch, for example.

Thomas

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
#5 lore raw

PATCH 1/5 qt5: bump to 5.1.1

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Thomas" == Thomas Petazzoni <[email protected]> writes:

Thomas> Dear Arnout Vandecappelle,

Quoted 1 line · Arnout Vandecappelle
Quoted 1 line

Perhaps we could try upstreaming our mkspecs? Does that make sense?

Thomas> Our mkspecs is really meant to be a template that has to be tuned by Thomas> the qt5base.mk makefile, and therefore outside Buildroot itself I Thomas> believe its meaning is not so important. We could even carry it within Thomas> package/qt5/qt5base/ rather than as a patch, for example.

That might indeed be a good idea.

Signature · 1 line
Bye, Peter Korsgaard
#6 lore raw

PATCH 1/5 qt5: bump to 5.1.1

In reply to Thomas Petazzoni

Dear Thomas,

Quoted 7 lines · On Thursday 19 September 2013 21:55:14 Thomas Petazzoni wrote:
Quoted 1 line · Arnout Vandecappelle

Perhaps we could try upstreaming our mkspecs? Does that make sense?

Our mkspecs is really meant to be a template that has to be tuned by the qt5base.mk makefile, and therefore outside Buildroot itself I believe its meaning is not so important. We could even carry it within package/qt5/qt5base/ rather than as a patch, for example.

I think, this part could be simplified by using a mkspec file for "Generic Linux" device which could be accepted by upstream. In order to customize variables we can use -device-option argument. As you mention in a comment in qt5base.mk, this option has some problems. I sent[1] a patch to upstream which fixes a problem when the value contains an equal sign.

I am also waiting[2] for a reply from upstream about the linux-generic-g++ device suggestion.

I have tried this method on my local machine and it seems working.

[1] https://bugreports.qt-project.org/browse/QTBUG-33584
[2] https://bugreports.qt-project.org/browse/QTBUG-33586

---
Fatih

#7 lore raw

PATCH 1/5 qt5: bump to 5.1.1

Dear Fatih A??c?,

Quoted 6 lines · Fatih A??c?

I think, this part could be simplified by using a mkspec file for "Generic Linux" device which could be accepted by upstream. In order to customize variables we can use -device-option argument. As you mention in a comment in qt5base.mk, this option has some problems. I sent[1] a patch to upstream which fixes a problem when the value contains an equal sign.

Quoted 7 lines

I am also waiting[2] for a reply from upstream about the linux-generic-g++ device suggestion.

I have tried this method on my local machine and it seems working.

[1] https://bugreports.qt-project.org/browse/QTBUG-33584
[2] https://bugreports.qt-project.org/browse/QTBUG-33586

Great, this definitely looks like a nice solution! I'm happy to see that you have taken the time to dive into Qt5 to polish all these little problems, this is really cool. Looking forward to seeing more progress on these things.

Thanks!

Thomas

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