[Buildroot] [PATCH v6 21/28] package/python-sip: Fix qmake usage for per-package build

[Andreas Naumann]

Hi Thomas,

On 11.03.20 23:30, Thomas Petazzoni wrote:
> Hello,
> 
> On Mon, 17 Feb 2020 22:23:43 +0100
> Andreas Naumann <anaumann at ultratronik.de> wrote:
> 
>> Since python-sip is not built by the qmake infra, we need to explicitly
>> integrate the qt.conf fixup needed for successful per-package build and
>> consistently use qmake via QT5_QMAKE variable.
>>
>> Signed-off-by: Andreas Naumann <anaumann at ultratronik.de>
>> ---
>>   package/python-sip/python-sip.mk | 3 ++-
>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/package/python-sip/python-sip.mk b/package/python-sip/python-sip.mk
>> index 1d6c205396..7ae2d43e1e 100644
>> --- a/package/python-sip/python-sip.mk
>> +++ b/package/python-sip/python-sip.mk
>> @@ -38,6 +38,7 @@ define HOST_PYTHON_SIP_INSTALL_CMDS
>>   endef
>>   
>>   define PYTHON_SIP_CONFIGURE_CMDS
>> +	$(QT5_QT_CONF_FIXUP)
>>   	(cd $(@D); \
>>   		$(TARGET_MAKE_ENV) $(TARGET_CONFIGURE_OPTS) $(HOST_DIR)/bin/python configure.py \
>>   			--bindir $(TARGET_DIR)/usr/bin \
>> @@ -46,7 +47,7 @@ define PYTHON_SIP_CONFIGURE_CMDS
>>   			--sipdir $(TARGET_DIR)/usr/share/sip \
>>   			--sysroot $(STAGING_DIR)/usr \
>>   			--use-qmake && \
>> -		$(HOST_DIR)/bin/qmake)
>> +		$(QT5_QMAKE))
> 
> Isn't this change a fix, independent from the per-package support ?

The QT5_QT_CONF_FIXUP is needed so python-sip doesnt litter qt5base 
during install, see patch 19. So no, this is per-package specific.

> 
> QT5_QMAKE is defined as follows:
> 
> QT5_QMAKE = $(HOST_DIR)/bin/qmake -spec devices/linux-buildroot-g++
> 
> So it means that python-sip was not using the
> devices/linux-buildroot-g++ spec file ? If that's the case, then this

Well, our qmake QMAKE_XSPEC is set to devices/linux-buildroot-g++ per 
default anyway so there was no problem in not explicitly setting it.
We do however need to use QT5_QMAKE to fix the problems described in 
patch 20.


best regards,
Andreas



> is something that should be fixed separately from the per-package
> support I believe.
> 
> Best regards,
> 
> Thomas
>