From: Andreas Oberritter <obi@opendreambox.org>
To: Andre McCurdy <armccurdy@gmail.com>,
openembedded-devel@lists.openembedded.org
Subject: Re: [meta-qt5][PATCH 3/3] qtwebkit: add packageconfig for gstreamer, qtlocation and qtsensors
Date: Fri, 04 Apr 2014 22:37:30 +0200 [thread overview]
Message-ID: <533F180A.80900@opendreambox.org> (raw)
In-Reply-To: <CAJ86T=XJ4XN5bDs4yKgUS7DFSkp9jSti7cKrDCYrcHRDyOtA6g@mail.gmail.com>
Hello Andre,
On 04.04.2014 05:34, Andre McCurdy wrote:
>> On Fri, Feb 28, 2014 at 12:15 PM, Andreas Oberritter
>> <obi@opendreambox.org> wrote:
>>> Use gstreamer for 1.x, gstreamer010 for 0.10.x.
>>>
>>> Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
>>> ---
>>> recipes-qt/qt5/qtwebkit.inc | 24 +++++++++++++++++++++++-
>>> 1 file changed, 23 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/recipes-qt/qt5/qtwebkit.inc b/recipes-qt/qt5/qtwebkit.inc
>>> index c35504e..90bd981 100644
>>> --- a/recipes-qt/qt5/qtwebkit.inc
>>> +++ b/recipes-qt/qt5/qtwebkit.inc
>>> @@ -5,7 +5,29 @@ LIC_FILES_CHKSUM = "file://Source/WebCore/rendering/RenderApplet.h;endline=22;md
>>> file://Source/WebKit/gtk/webkit/webkit.h;endline=21;md5=b4fbe9f4a944f1d071dba1d2c76b3351 \
>>> file://Source/JavaScriptCore/parser/Parser.h;endline=21;md5=bd69f72183a7af673863f057576e21ee"
>>>
>>> -DEPENDS += "qtdeclarative qtlocation qtsensors icu ruby-native sqlite3 glib-2.0 gstreamer1.0 gstreamer1.0-plugins-base libxslt"
>>> +DEPENDS += "qtbase qtdeclarative icu ruby-native sqlite3 glib-2.0 libxslt"
>>> +
>>> +PACKAGECONFIG ??= "gstreamer qtlocation qtsensors"
>>> +PACKAGECONFIG[gstreamer] = ",,gstreamer1.0 gstreamer1.0-plugins-base"
>>> +PACKAGECONFIG[gstreamer010] = ",,gstreamer gst-plugins-base"
>>> +PACKAGECONFIG[qtlocation] = ",,qtlocation"
>>> +PACKAGECONFIG[qtsensors] = ",,qtsensors"
>>> +
>>> +do_configure_prepend() {
>>> + # disable gstreamer-1.0 test if it isn't enabled by PACKAGECONFIG
>>> + sed -e 's/\s\(packagesExist(".*\<gstreamer-1.0\>.*")\)/ OE_GSTREAMER_ENABLED:\1/' -i ${S}/Tools/qmake/mkspecs/features/features.prf
>>> + # disable gstreamer-0.10 test if it isn't enabled by PACKAGECONFIG
>>> + sed -e 's/\s\(packagesExist(".*\<gstreamer-0.10\>.*")\)/ OE_GSTREAMER010_ENABLED:\1/' -i ${S}/Tools/qmake/mkspecs/features/features.prf
>>> + # disable qtlocation test if it isn't enabled by PACKAGECONFIG
>>> + sed -e 's/\s\(qtHaveModule(positioning)\)/ OE_QTLOCATION_ENABLED:\1/' -i ${S}/Tools/qmake/mkspecs/features/features.prf
>>> + # disable qtsensors test if it isn't enabled by PACKAGECONFIG
>>> + sed -e 's/\s\(qtHaveModule(sensors)\)/ OE_QTSENSORS_ENABLED:\1/' -i ${S}/Tools/qmake/mkspecs/features/features.prf
>>> +}
>>> +
>>> +EXTRA_QMAKEVARS_PRE += "${@base_contains('PACKAGECONFIG', 'gstreamer', 'CONFIG+=OE_GSTREAMER_ENABLED', '', d)}"
>>> +EXTRA_QMAKEVARS_PRE += "${@base_contains('PACKAGECONFIG', 'gstreamer010', 'CONFIG+=OE_GSTREAMER010_ENABLED', '', d)}"
>>> +EXTRA_QMAKEVARS_PRE += "${@base_contains('PACKAGECONFIG', 'qtlocation', 'CONFIG+=OE_QTLOCATION_ENABLED', '', d)}"
>>> +EXTRA_QMAKEVARS_PRE += "${@base_contains('PACKAGECONFIG', 'qtsensors', 'CONFIG+=OE_QTSENSORS_ENABLED', '', d)}"
>>>
>>> # qtwebkit gets terribly big when linking with all debug info, disable by default
>>> QTWEBKIT_DEBUG = "QMAKE_CFLAGS+=-g0 QMAKE_CXXFLAGS+=-g0"
>>> --
>>
>> With this change I'm seeing that the CONFIG+= options from the qmake
>> command line are being added to CONFIG after features.prf has already
>> been evaluated. Therefore OE_GSTREAMER_ENABLED etc evaluates as false
>> and qtwebkit is always configured without video support.
>>
>> From the configure log:
>>
>>> ! Missing GStreamer or QtMultimedia, disabling HTML5 media element support
>>
>> From the qmake debug log:
>>
>>> DEBUG 1: ... features.prf:106: condition "OE_GSTREAMER_ENABLED" is false (around line 5480 of the log)
>>> ...
>>> DEBUG 1: (command line):65535: CONFIG := ... OE_GSTREAMER_ENABLED (around line 5802 of the log, just prior to starting on WebKit.pro).
>>
>> I'm testing with dora + master branch of meta-qt5.
>>
>
> Hi all,
>
> I'm still seeing problems since this commit. Even when these new
> options are enabled via PACKAGECONFIG, qtwebkit is being configured as
> if they were all disabled.
>
> The issue seems to be that the "CONFIG+=OE_xxx_ENABLED" options on the
> qmake command line don't work as intended. The options are added to
> CONFIG _after_ features.prf is parsed, so testing them in features.prf
> always evaluates to false.
>
> Symptoms are "disabling HTML5 media element support" showing up in the
> qtwebkit configure log and libQt5WebKit.so.5.2.1 not having any
> runtime dependency on libgstreamer, etc.
>
> Is anyone else seeing the same thing ?
thanks for your reminder. After your first mail I got distracted and
forgot about looking into this.
I was able to reproduce it and just submitted a patch to the mailing
list. I build-tested it with PACKAGECONFIG = "", "gstreamer" and
"gstreamer qtlocation" on dora.
Regards,
Andreas
next prev parent reply other threads:[~2014-04-04 20:37 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-02-28 20:15 [meta-qt5][PATCH 1/3] qtbase/qtwebkit: drop ICU variable Andreas Oberritter
2014-02-28 20:15 ` [meta-qt5][PATCH 2/3] qtdeclarative: drop qtsvg dependency, add packageconfig for qtxmlpatterns Andreas Oberritter
2014-02-28 20:15 ` [meta-qt5][PATCH 3/3] qtwebkit: add packageconfig for gstreamer, qtlocation and qtsensors Andreas Oberritter
2014-03-28 6:16 ` Andre McCurdy
2014-04-04 3:34 ` Andre McCurdy
2014-04-04 20:37 ` Andreas Oberritter [this message]
2014-04-21 11:43 ` Martin Jansa
2014-02-28 21:28 ` [meta-qt5][PATCH 1/3] qtbase/qtwebkit: drop ICU variable Gary Thomas
2014-02-28 22:29 ` Andreas Oberritter
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=533F180A.80900@opendreambox.org \
--to=obi@opendreambox.org \
--cc=armccurdy@gmail.com \
--cc=openembedded-devel@lists.openembedded.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.