Error building qt4e-base-image with angstrom-2010.x

Error building qt4e-base-image with angstrom-2010.x

[Marco Cavallini]

Hi,
I'm facing to a problem building qt4e-base-image with angstrom-2010.x on
a Debian lenny host, is it normal?
http://pastebin.com/6SnFSYs6
I suspect a Debian lenny issue with new OE stuff.

Any hint would be appreciated.
TIA
--
Marco

Re: Error building qt4e-base-image with angstrom-2010.x

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 691 bytes --]

Dear Marco,


Am Montag, den 30.05.2011, 10:18 +0200 schrieb Marco Cavallini:

> I'm facing to a problem building qt4e-base-image with angstrom-2010.x on
> a Debian lenny host, is it normal?
> http://pastebin.com/6SnFSYs6

please always paste the logs into the message so that people do not need
to open a browser, can comment inline and have everything in the archive
too.

> I suspect a Debian lenny issue with new OE stuff.

I would not think so, because the tool chain used for compiling is build
extra (native/sysroot).

> Any hint would be appreciated.

Does it build with another distribution or tool chain? Did the build
ever work for you?


Thanks,

Paul

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

Re: Error building qt4e-base-image with angstrom-2010.x

[Marco Cavallini]

[-- Attachment #1: Type: text/plain, Size: 4323 bytes --]

Paul Menzel ha scritto, Il 30/05/2011 10:35:
> Dear Marco,
> 
> 
> Am Montag, den 30.05.2011, 10:18 +0200 schrieb Marco Cavallini:
> 
>> I'm facing to a problem building qt4e-base-image with angstrom-2010.x on
>> a Debian lenny host, is it normal?
>> http://pastebin.com/6SnFSYs6
> 
> please always paste the logs into the message so that people do not need
> to open a browser, can comment inline and have everything in the archive
> too.

See attachment.


>> I suspect a Debian lenny issue with new OE stuff.
> 
> I would not think so, because the tool chain used for compiling is build
> extra (native/sysroot).
> 
>> Any hint would be appreciated.
> 
> Does it build with another distribution or tool chain? Did the build
> ever work for you?
> 

I built smoothly until March 2011, today I'm unable to build using
angstrom-2010.x nor kaeilos


TIA
--
Marco

d-emb-arm/qrc_WebKit.o .rcc/release-shared-emb-arm/qrc_WebKit.cpp
| In file included from ../../../../include/QtCore/qstring.h:1:0,
|                  from ../WebKit/qt/Api/qwebsettings.h:25,
|                  from ../WebKit/qt/Api/qwebpage.h:24,
|                  from ../WebKit/qt/Api/qwebpage.cpp:23:
| ../../../../include/QtCore/../../src/corelib/tools/qstring.h: At global scope:
| ../../../../include/QtCore/../../src/corelib/tools/qstring.h:187:17: note: the mangling of 'va_list' has changed in GCC 4.4
| In file included from ../../../../include/QtCore/qstring.h:1:0,
|                  from .moc/release-shared-emb-arm/../../../WebKit/qt/Api/qwebsettings.h:25,
|                  from .moc/release-shared-emb-arm/../../../WebKit/qt/Api/qwebpage.h:24,
|                  from .moc/release-shared-emb-arm/../../../WebKit/qt/Api/qwebinspector.h:24,
|                  from .moc/release-shared-emb-arm/moc_qwebinspector.cpp:10:
| ../../../../include/QtCore/../../src/corelib/tools/qstring.h:187:17: note: the mangling of 'va_list' has changed in GCC 4.4
| rm -f libQtWebKitE.so.4.7.3 libQtWebKitE.so libQtWebKitE.so.4 libQtWebKitE.so.4.7
| arm-angstrom-linux-gnueabi-g++ -march=armv5te -mtune=arm926ej-s -mthumb-interwork -mno-thumb --sysroot=/home/koan/devel/build/angstrom/tmp/sysroots/armv5te-angstrom-linux-gnueabi -Wl,-rpath-link,/home/koan/devel/build/angstrom/tmp/sysroots/armv5te-angstrom-linux-gnueabi/usr/lib -Wl,-O1 -Wl,--hash-style=gnu -Wl,--no-undefined -Wl,--gc-sections -Wl,--no-undefined -Wl,-rpath-link,/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/qt4-embedded-4.7.3-r33.1/qt-everywhere-opensource-src-4.7.3/lib -Wl,-rpath-link,../JavaScriptCore/release -Wl,-rpath-link,/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/qt4-embedded-4.7.3-r33.1/qt-everywhere-opensource-src-4.7.3/lib -shared -Wl,-Bsymbolic-functions -Wl,-soname,libQtWebKitE.so.4

...

| /home/koan/devel/build/angstrom/tmp/sysroots/x86_64-linux/usr/armv5te/lib/gcc/arm-angstrom-linux-gnueabi/4.5.3/../../../../arm-angstrom-linux-gnueabi/bin/ld: BFD (GNU Binutils) 2.20.1.20100303 internal error, aborting at /home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/binutils-cross-2.20.1-r14.4/binutils-2.20.1/bfd/merge.c line 876 in _bfd_merged_section_offset
|
| /home/koan/devel/build/angstrom/tmp/sysroots/x86_64-linux/usr/armv5te/lib/gcc/arm-angstrom-linux-gnueabi/4.5.3/../../../../arm-angstrom-linux-gnueabi/bin/ld: Please report this bug.
|
| collect2: ld returned 1 exit status
| make[1]: *** [../../../../lib/libQtWebKitE.so.4.7.3] Error 1
| make[1]: Leaving directory `/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/qt4-embedded-4.7.3-r33.1/qt-everywhere-opensource-src-4.7.3/src/3rdparty/webkit/WebCore'
| make: *** [sub-webkit-make_default-ordered] Error 2
| FATAL: oe_runmake failed
| ERROR: Function do_compile failed
NOTE: package qt4-embedded-4.7.3-r33.1: task do_compile: Failed
ERROR: TaskFailed event exception, aborting
ERROR: Build of /home/koan/devel/openembedded/recipes/qt4/qt4-embedded_4.7.3.bb do_compile failed
ERROR: Task 400 (/home/koan/devel/openembedded/recipes/qt4/qt4-embedded_4.7.3.bb, do_compile) failed with 256
ERROR: '/home/koan/devel/openembedded/recipes/qt4/qt4-embedded_4.7.3.bb' failed
ERROR: '/home/koan/devel/openembedded/recipes/qt4/qt4-embedded_4.7.3.bb' failed

Re: Error building qt4e-base-image with angstrom-2010.x

[Henning Heinold]

Hi,

in March you didnt build qt 4.7.3 for sure.

Bye Henning

Re: Error building qt4e-base-image with angstrom-2010.x

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 30-05-11 10:18, Marco Cavallini wrote:
> Hi,
> I'm facing to a problem building qt4e-base-image with angstrom-2010.x on
> a Debian lenny host, is it normal?
> http://pastebin.com/6SnFSYs6
> I suspect a Debian lenny issue with new OE stuff.

I just tried this on my debian box (dunno which version):

NOTE: package qt4-embedded-4.7.3-r33.1: task do_build: Succeeded
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFN46w3MkyGM64RGpERAobXAJ9M6ji5iJTvdewhuT61ZM+9QpKVhgCfZ4TG
QICHAHYa8VRIkdRG+FHgGlc=
=aD4m
-----END PGP SIGNATURE-----

Re: Error building qt4e-base-image with angstrom-2010.x

[Marco Cavallini]

Koen Kooi ha scritto, Il 30/05/2011 16:39:
> On 30-05-11 10:18, Marco Cavallini wrote:
>> Hi,
>> I'm facing to a problem building qt4e-base-image with angstrom-2010.x on
>> a Debian lenny host, is it normal?
>> http://pastebin.com/6SnFSYs6
>> I suspect a Debian lenny issue with new OE stuff.
> 
> I just tried this on my debian box (dunno which version):
> 
> NOTE: package qt4-embedded-4.7.3-r33.1: task do_build: Succeeded

Thank you Koen for your confirmation test.

Re: Error building qt4e-base-image with angstrom-2010.x

[Marco Cavallini]

Koen Kooi ha scritto, Il 30/05/2011 16:39:
> On 30-05-11 10:18, Marco Cavallini wrote:
>> Hi,
>> I'm facing to a problem building qt4e-base-image with angstrom-2010.x on
>> a Debian lenny host, is it normal?
>> http://pastebin.com/6SnFSYs6
>> I suspect a Debian lenny issue with new OE stuff.
> 
> I just tried this on my debian box (dunno which version):
> 
> NOTE: package qt4-embedded-4.7.3-r33.1: task do_build: Succeeded


Which machine did you set?


BTW looks like I have a major problem with gcc 4.4 and armel - the ABI
of va_list changed (for c++ libraries)
I see the same error with 'micro' distro.
My machine is MACHINE ?= "ronetix-pm9263"


TIA
--
Marco

Re: Error building qt4e-base-image with angstrom-2010.x

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 31-05-11 08:59, Marco Cavallini wrote:
> Koen Kooi ha scritto, Il 30/05/2011 16:39:
>> On 30-05-11 10:18, Marco Cavallini wrote:
>>> Hi,
>>> I'm facing to a problem building qt4e-base-image with angstrom-2010.x on
>>> a Debian lenny host, is it normal?
>>> http://pastebin.com/6SnFSYs6
>>> I suspect a Debian lenny issue with new OE stuff.
>>
>> I just tried this on my debian box (dunno which version):
>>
>> NOTE: package qt4-embedded-4.7.3-r33.1: task do_build: Succeeded
> 
> 
> Which machine did you set?

hawkboard
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFN5Jv7MkyGM64RGpERAiszAJ0Z//I1f4PlWAGTZdilcGiMJp+FCwCaA4Nt
dMGAg6YMgCJg9GNm0lF/EMY=
=dcNK
-----END PGP SIGNATURE-----

Re: Error building qt4e-base-image with angstrom-2010.x

[Phil Blundell]

On Tue, 2011-05-31 at 08:59 +0200, Marco Cavallini wrote:
> BTW looks like I have a major problem with gcc 4.4 and armel - the ABI
> of va_list changed (for c++ libraries)

Why is this a problem? 

p.

Re: Error building qt4e-base-image with angstrom-2010.x

[Marco Cavallini]

Phil Blundell ha scritto, Il 31/05/2011 15:25:
> On Tue, 2011-05-31 at 08:59 +0200, Marco Cavallini wrote:
>> BTW looks like I have a major problem with gcc 4.4 and armel - the ABI
>> of va_list changed (for c++ libraries)
> 
> Why is this a problem? 
> 
> p.
> 


Phil,
looks like the problem is gcc-4.3 on the Host side (Debian Lenny x86_64)
and besides I wonder if I'm the only one OE user using this distro I'm
doing a backup of the whole system and planned to upgrade to Squeeze
tomorrow.

This is the output:

| gcc  -c
-isystem/home/koan/devel/build/angstrom/tmp/sysroots/x86_64-linux/usr/include
-O2 -g -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE  -W -Wall -Wwrite-strings
-Wcast-qual -Wstrict-prototypes -Wmissing-prototypes
-Wmissing-format-attribute -pedantic -Wno-long-long -Wno-variadic-macros
-Wno-overlength-strings -Wold-style-definition -Wc++-compat
-DHAVE_CONFIG_H -I. -I.
-I/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc
-I/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/.
-I/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/../include
-I/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/../libcpp/include
-I/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/../libdecnumber
-I/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/../libdecnumber/dpd
-I../libdecnumber   -I/usr/include/libelf
/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/dwarf2out.c
-o dwarf2out.o
|
/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/config/arm/neon.md:
In function 'recog_57':
|
/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/config/arm/neon.md:5730:
internal compiler error: Segmentation fault
| Please submit a full bug report,
| with preprocessed source if appropriate.
| See <file:///usr/share/doc/gcc-4.3/README.Bugs> for instructions.
| make[2]: *** [insn-recog.o] Error 1
| make[2]: *** Waiting for unfinished jobs....
|
/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/dwarf2out.c:6170:
warning: 'is_redundant_typedef' declared inline after being called
|
/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/dwarf2out.c:6170:
warning: previous declaration of 'is_redundant_typedef' was here
| rm gcov.pod gfdl.pod cpp.pod gcc.pod fsf-funding.pod
| make[2]: Leaving directory
`/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/build.x86_64-linux.arm-angstrom-linux-gnueabi/gcc'
| make[1]: *** [all-gcc] Error 2
| make[1]: Leaving directory
`/home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/build.x86_64-linux.arm-angstrom-linux-gnueabi'
| make: *** [all] Error 2
| FATAL: oe_runmake failed
| ERROR: Function do_compile failed
NOTE: package gcc-cross-initial-4.5-r36.1+svnr170880: task do_compile:
Failed
ERROR: TaskFailed event exception, aborting
ERROR: Build of
/home/koan/devel/openembedded/recipes/gcc/gcc-cross-initial_4.5.bb
do_compile failed
ERROR: Task 2258
(/home/koan/devel/openembedded/recipes/gcc/gcc-cross-initial_4.5.bb,
do_compile) failed with 256
Waiting for 1 active tasks to finish:
1: util-linux-ng-native-2.18-r36.0 do_install (pid 4670)
ERROR:
'/home/koan/devel/openembedded/recipes/gcc/gcc-cross-initial_4.5.bb' failed
NOTE: package util-linux-ng-native-2.18-r36.0: task do_install: Succeeded
ERROR:
'/home/koan/devel/openembedded/recipes/gcc/gcc-cross-initial_4.5.bb' failed
ERROR:
'/home/koan/devel/openembedded/recipes/gcc/gcc-cross-initial_4.5.bb' failed

Re: Error building qt4e-base-image with angstrom-2010.x

[Phil Blundell]

On Tue, 2011-05-31 at 15:49 +0200, Marco Cavallini wrote:
> Phil Blundell ha scritto, Il 31/05/2011 15:25:
> > On Tue, 2011-05-31 at 08:59 +0200, Marco Cavallini wrote:
> >> BTW looks like I have a major problem with gcc 4.4 and armel - the ABI
> >> of va_list changed (for c++ libraries)
> > 
> > Why is this a problem? 
>
> Phil,
> looks like the problem is gcc-4.3 on the Host side (Debian Lenny x86_64)
> and besides I wonder if I'm the only one OE user using this distro I'm
> doing a backup of the whole system and planned to upgrade to Squeeze
> tomorrow.
> 
> This is the output:
> 
> /home/koan/devel/build/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-initial-4.5-r36.1+svnr170880/gcc-4_5-branch/gcc/config/arm/neon.md:5730:
> internal compiler error: Segmentation fault
> | Please submit a full bug report,
> | with preprocessed source if appropriate.
> | See <file:///usr/share/doc/gcc-4.3/README.Bugs> for instructions.

That certainly does indicate a bug in your host gcc (if it's
reproducible in the same place) or a broken computer (if it occurs at
different, random places).  Assuming the first, you might be able to
work around it by building the cross gcc with less optimisation, but if
you're going to upgrade to squeeze anyway then that probably will solve
the problem too.

This doesn't seem to have anything to do with the thing you originally
mentioned (ARM va_list ABI change in gcc 4.4) though.

p.