Subversion-1.7.0 and arm/thumb work arounds
cover Subversion-1.7.0 and arm/thumb work arounds
Hi,
arm mode was forced without PR bump, because if it was failing for someone he will notice.
Be carefull with using subversion-1.7.0 (read the commit message).
Pulseaudio-1.1 was released, so hopefully this work arounds won't be needed there or they could be resolved better with upgrade.
Regards,
The following changes since commit 99da9a4e65f9dffb04efc3ad60125194c476d6b3:
distro-tracking-fields: update fields for tzdata and gst-plugins-good (2011-10-20 13:07:16 +0100)
are available in the git repository at:
git://git.openembedded.org/openembedded-core-contrib jansa/pull2
http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=jansa/pull2
Martin Jansa (8):
pulseaudio-0.9.23: inherit perlnative to work around build on host
without XML/Parser.pm
libatomics-ops: force ARM mode
pulseaudio-0.9.23: force ARM mode
webkit-gtk: force arm mode to work around binutils segfault
apr: add native support
neon: add native support
apr-util: add native support
subversion: add 1.7.0 with native support and negative D_P for now
.../subversion/subversion-1.7.0/libtool2.patch | 15 ++++++++ .../subversion/subversion_1.7.0.bb | 37 ++++++++++++++++++++ .../pulseaudio/libatomics-ops_1.2.bb | 2 + .../pulseaudio/pulseaudio_0.9.23.bb | 3 +- meta/recipes-sato/webkit/webkit-gtk_svn.bb | 17 +++++++++ meta/recipes-support/apr/apr-util_1.3.12.bb | 8 ++++ meta/recipes-support/apr/apr_1.4.5.bb | 2 + meta/recipes-support/neon/neon_0.29.5.bb | 3 ++ 8 files changed, 86 insertions(+), 1 deletions(-) create mode 100644 meta/recipes-devtools/subversion/subversion-1.7.0/libtool2.patch create mode 100644 meta/recipes-devtools/subversion/subversion_1.7.0.bb
Signature · 1 line
1.7.7
PATCH 1/8 pulseaudio-0.9.23: inherit perlnative to work around build on host without XML/Parser.pm
--- .../pulseaudio/pulseaudio_0.9.23.bb | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +1 -1
@@ -4,7 +4,7 @@ PR = "r5" DEPENDS += "gdbm speex" -inherit gettext +inherit gettext perlnative SRC_URI = "http://freedesktop.org/software/pulseaudio/releases/pulseaudio-${PV}.tar.gz \ file://buildfix.patch \
Signature · 1 line
1.7.7
PATCH 2/8 libatomics-ops: force ARM mode
* otherwise ie spitz (armv5te) build fails with:
| make[3]: Entering directory `/OE/shr-core/tmp/work/armv5te-oe-linux-gnueabi/libatomics-ops-1.2-r5/libatomic_ops-1.2/src'
| arm-oe-linux-gnueabi-gcc -march=armv5te -mthumb -mthumb-interwork -mtune=xscale --sysroot=/OE/shr-core/tmp/sysroots/spitz -DHAVE_CONFIG_H -I. -fPIC -O
2 -pipe -g -feliminate-unused-debug-types -DNDEBUG -c atomic_ops.c
| arm-oe-linux-gnueabi-gcc -march=armv5te -mthumb -mthumb-interwork -mtune=xscale --sysroot=/OE/shr-core/tmp/sysroots/spitz -DHAVE_CONFIG_H -I. -fPIC -O
2 -pipe -g -feliminate-unused-debug-types -DNDEBUG -c atomic_ops_stack.c
| arm-oe-linux-gnueabi-gcc -march=armv5te -mthumb -mthumb-interwork -mtune=xscale --sysroot=/OE/shr-core/tmp/sysroots/spitz -DHAVE_CONFIG_H -I. -fPIC -O
2 -pipe -g -feliminate-unused-debug-types -DNDEBUG -c atomic_ops_malloc.c
| atomic_ops_malloc.c: In function 'msb':
| atomic_ops_malloc.c:223:2: warning: right shift count >= width of type [enabled by default]
| rm -f libatomic_ops_gpl.a
| ar cru libatomic_ops_gpl.a atomic_ops_stack.o atomic_ops_malloc.o
| arm-oe-linux-gnueabi-ranlib libatomic_ops_gpl.a
| {standard input}: Assembler messages:
| {standard input}:286: Error: selected processor does not support Thumb mode `swp r1,r2,[r3]'
| {standard input}:329: Error: selected processor does not support Thumb mode `swp r0,r1,[r3]'--- .../pulseaudio/libatomics-ops_1.2.bb | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
meta/recipes-multimedia/pulseaudio/libatomics-ops_1.2.bb +2 -0
@@ -20,4 +20,6 @@ S = "${WORKDIR}/libatomic_ops-${PV}" ALLOW_EMPTY_${PN} = "1" +ARM_INSTRUCTION_SET = "arm" + inherit autotools pkgconfig
Signature · 1 line
1.7.7
Re: [PATCH 2/8] libatomics-ops: force ARM mode
Hi,
hm we should think of reworking this recipe now. Because since gcc 4.5 pulseaudio for arm can use the gcc internal atomicstuff and in oe-core and meta-oe we have 4.5 or 4.6 only. The lib is
only needed for mips and it is still the old release, on cvs is a much better version, which supports thumb too, if remember correctly.
Bye Henning
Re: [PATCH 2/8] libatomics-ops: force ARM mode
Quoted 8 lines · Henning Heinold
Hi,
hm we should think of reworking this recipe now. Because since gcc 4.5 pulseaudio for arm can use the gcc internal atomicstuff and in oe-core and meta-oe we have 4.5 or 4.6 only. The lib is
only needed for mips and it is still the old release, on cvs is a much better version, which supports thumb too, if remember correctly.
Agreed, I've added this only because pulseaudio is now needed for gst-plugins and I wasn't able to build any armv[45]t images.
But because I don't know much about pulseaudio or libatomics-ops I was hoping that someone who maintains them will fix it properly and this work around will be needed only temporary.
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
Re: [PATCH 2/8] libatomics-ops: force ARM mode
Quoted 10 lines · Martin Jansa
Quoted 8 lines · Henning Heinold
Hi,
hm we should think of reworking this recipe now. Because since gcc 4.5 pulseaudio for arm can use the gcc internal atomicstuff and in oe-core and meta-oe we have 4.5 or 4.6 only. The lib is
only needed for mips and it is still the old release, on cvs is a much better version, which supports thumb too, if remember correctly.
Quoted 6 lines
Agreed, I've added this only because pulseaudio is now needed for gst-plugins and I wasn't able to build any armv[45]t images.
But because I don't know much about pulseaudio or libatomics-ops I was hoping that someone who maintains them will fix it properly and this work around will be needed only temporary.
Would you have cycles to create a git recipe for it the git tree is here
https://github.com/ivmai/libatomic_ops/
Quoted 10 lines
--
Martin 'JaMa' Jansa jabber: [email protected]_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [PATCH 2/8] libatomics-ops: force ARM mode
Quoted 14 lines · Khem Raj
Quoted 13 lines · Martin Jansa
Quoted 8 lines · Henning Heinold
Hi,
hm we should think of reworking this recipe now. Because since gcc 4.5 pulseaudio for arm can use the gcc internal atomicstuff and in oe-core and meta-oe we have 4.5 or 4.6 only. The lib is
only needed for mips and it is still the old release, on cvs is a much better version, which supports thumb too, if remember correctly.Agreed, I've added this only because pulseaudio is now needed for gst-plugins and I wasn't able to build any armv[45]t images.
Quoted 6 lines
Quoted 3 lines
But because I don't know much about pulseaudio or libatomics-ops I was hoping that someone who maintains them will fix it properly and this work around will be needed only temporary.
Would you have cycles to create a git recipe for it the git tree is here
https://github.com/ivmai/libatomic_ops/
Sorry, I'm not going to spend more of my free time on stuff I've never used on my phones at least as long as we have broken stuff which we're using almost daily.
But I can keep this work around or .bbappend dropping whole pulseaudio in SHR until someone fixes it properly or until I fix all other issues we have.
Regards,
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
Re: [PATCH 2/8] libatomics-ops: force ARM mode
In reply to Martin JansaQuoted 27 lines · Martin Jansa
* otherwise ie spitz (armv5te) build fails with:
| make[3]: Entering directory `/OE/shr-core/tmp/work/armv5te-oe-linux-gnueabi/libatomics-ops-1.2-r5/libatomic_ops-1.2/src'
| arm-oe-linux-gnueabi-gcc -march=armv5te -mthumb -mthumb-interwork -mtune=xscale --sysroot=/OE/shr-core/tmp/sysroots/spitz -DHAVE_CONFIG_H -I. -fPIC -O
2 -pipe -g -feliminate-unused-debug-types -DNDEBUG -catomic_ops.c
| arm-oe-linux-gnueabi-gcc -march=armv5te -mthumb -mthumb-interwork -mtune=xscale --sysroot=/OE/shr-core/tmp/sysroots/spitz -DHAVE_CONFIG_H -I. -fPIC -O
2 -pipe -g -feliminate-unused-debug-types -DNDEBUG -catomic_ops_stack.c
| arm-oe-linux-gnueabi-gcc -march=armv5te -mthumb -mthumb-interwork -mtune=xscale --sysroot=/OE/shr-core/tmp/sysroots/spitz -DHAVE_CONFIG_H -I. -fPIC -O
2 -pipe -g -feliminate-unused-debug-types -DNDEBUG -catomic_ops_malloc.c| atomic_ops_malloc.c: In function 'msb': | atomic_ops_malloc.c:223:2: warning: right shift count >= width of type [enabled by default] | rm -f libatomic_ops_gpl.a | ar cru libatomic_ops_gpl.a atomic_ops_stack.o atomic_ops_malloc.o | arm-oe-linux-gnueabi-ranlib libatomic_ops_gpl.a | {standard input}: Assembler messages: | {standard input}:286: Error: selected processor does not support Thumb mode `swp r1,r2,[r3]' | {standard input}:329: Error: selected processor does not support Thumb mode `swp r0,r1,[r3]'Signed-off-by Martin Jansa <[email protected]>--- .../pulseaudio/libatomics-ops_1.2.bb | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)meta/recipes-multimedia/pulseaudio/libatomics-ops_1.2.bb +0
-0
meta/recipes-multimedia/pulseaudio/libatomics-ops_1.2.bb S = "${WORKDIR}/libatomic_ops-${PV}" · line 25
Quoted 6 lines
meta/recipes-multimedia/pulseaudio/libatomics-ops_1.2.bb +2
-0@@ -20,4 +20,6 @@ S = "${WORKDIR}/libatomic_ops-${PV}" ALLOW_EMPTY_${PN} = "1" +ARM_INSTRUCTION_SET = "arm" + inherit autotools pkgconfig
Wouldn't it be better to fix libatomic-ops to have proper thumb mode assembly, then fall back to non-thumb assembly?
I swear I had a patch to do this in the past, but I'm currently not able to find it.
--Mark
PATCH 3/8 pulseaudio-0.9.23: force ARM mode
In reply to Martin Jansa--- .../pulseaudio/pulseaudio_0.9.23.bb | 1 + 1 files changed, 1 insertions(+), 0 deletions(-)
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +1 -0
@@ -23,3 +23,4 @@ do_compile_prepend() { cp ${STAGING_LIBDIR}/libltdl* ${S}/libltdl } +ARM_INSTRUCTION_SET = "arm"
Signature · 1 line
1.7.7
Re: [PATCH 3/8] pulseaudio-0.9.23: force ARM mode
Op 21 okt. 2011, om 10:17 heeft Martin Jansa het volgende geschreven:
Quoted 10 lines
Signed-off-by Martin Jansa <[email protected]>---
.../pulseaudio/pulseaudio_0.9.23.bb | 1 +1 files changed, 1 insertions(+), 0 deletions(-)meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +0
-0
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb do_compile_prepend() { · line 26
Quoted 4 lines
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +1
-0@@ -23,3 +23,4 @@ do_compile_prepend() { cp ${STAGING_LIBDIR}/libltdl* ${S}/libltdl } +ARM_INSTRUCTION_SET = "arm"
Missing PR bump
Re: [PATCH 3/8] pulseaudio-0.9.23: force ARM mode
In reply to Martin JansaQuoted 13 lines · Martin Jansa
Signed-off-by Martin Jansa <[email protected]>---
.../pulseaudio/pulseaudio_0.9.23.bb | 1 +1 files changed, 1 insertions(+), 0 deletions(-)meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +0
-0@@ -23,3 +23,4 @@ do_compile_prepend() {cp ${STAGING_LIBDIR}/libltdl* ${S}/libltdl
}
Quoted 1 line
+ARM_INSTRUCTION_SET = "arm"
not averse to patch but it would be nice to know what fails in thumb mode.
Quoted 9 lines
--
1.7.7_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [PATCH 3/8] pulseaudio-0.9.23: force ARM mode
Op 21 okt. 2011, om 17:40 heeft Khem Raj het volgende geschreven:
Quoted 14 lines · Martin Jansa
Quoted 13 lines
Signed-off-by Martin Jansa <[email protected]>--- .../pulseaudio/pulseaudio_0.9.23.bb | 1 + 1 files changed, 1 insertions(+), 0 deletions(-)meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +0
-0@@ -23,3 +23,4 @@ do_compile_prepend() { cp ${STAGING_LIBDIR}/libltdl* ${S}/libltdl }
Quoted 3 lines
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb do_compile_prepend() { · line 26
Quoted 1 line
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +1
-0+ARM_INSTRUCTION_SET = "arm"not averse to patch but it would be nice to know what fails in thumb mode.
And pulse is at 1.1 now, maybe upstream has fixed this issue.
Re: [PATCH 3/8] pulseaudio-0.9.23: force ARM mode
In reply to Khem RajQuoted 14 lines · Khem Raj
Quoted 13 lines · Martin Jansa
Signed-off-by Martin Jansa <[email protected]>---
.../pulseaudio/pulseaudio_0.9.23.bb | 1 +1 files changed, 1 insertions(+), 0 deletions(-)meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +0
-0@@ -23,3 +23,4 @@ do_compile_prepend() {cp ${STAGING_LIBDIR}/libltdl* ${S}/libltdl
}
Quoted 3 lines
Quoted 1 line
+ARM_INSTRUCTION_SET = "arm"
not averse to patch but it would be nice to know what fails in thumb mode.
selected processor does not support Thumb mode `swp r1,r2,[r3]'
Regards,
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
Re: [PATCH 3/8] pulseaudio-0.9.23: force ARM mode
Quoted 15 lines · Martin Jansa
Quoted 14 lines · Khem Raj
Quoted 13 lines · Martin Jansa
Signed-off-by Martin Jansa<[email protected]>--- .../pulseaudio/pulseaudio_0.9.23.bb | 1 + 1 files changed, 1 insertions(+), 0 deletions(-)meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +0
-0@@ -23,3 +23,4 @@ do_compile_prepend() { cp ${STAGING_LIBDIR}/libltdl* ${S}/libltdl }
Quoted 6 lines
Quoted 3 lines
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb do_compile_prepend() { · line 26
Quoted 1 line
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +1
-0+ARM_INSTRUCTION_SET = "arm"not averse to patch but it would be nice to know what fails in thumb mode.
selected processor does not support Thumb mode `swp r1,r2,[r3]'
As already asked, can you upgrade to 1.1 and verify if this problem still exists. Also if it does, it would be good to have this info in the commit message.
Thanks
Sau!
Quoted 9 lines
Regards,
_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [PATCH 3/8] pulseaudio-0.9.23: force ARM mode
In reply to Martin JansaQuoted 15 lines · Martin Jansa
Quoted 14 lines · Khem Raj
Quoted 13 lines · Martin Jansa
Signed-off-by Martin Jansa <[email protected]>--- .../pulseaudio/pulseaudio_0.9.23.bb | 1 + 1 files changed, 1 insertions(+), 0 deletions(-)meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +0
-0@@ -23,3 +23,4 @@ do_compile_prepend() { cp ${STAGING_LIBDIR}/libltdl* ${S}/libltdl }
Quoted 5 lines
Quoted 3 lines
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb do_compile_prepend() { · line 26
Quoted 1 line
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +1
-0+ARM_INSTRUCTION_SET = "arm"not averse to patch but it would be nice to know what fails in thumb mode.
selected processor does not support Thumb mode `swp r1,r2,[r3]'
If it's only that one error (presumably some kind of home-grown atomic operation) then let's just fix the code in question rather than forcing the whole thing to build in ARM-state. Or if the code is hard to fix, patch the makefile so that only the particular source files with the assembly in are built for ARM and let the rest stay as Thumb. Making the whole package be ARM is a bit of a heavyweight fix for that issue.
p.
Re: [PATCH 3/8] pulseaudio-0.9.23: force ARM mode
Quoted 20 lines · Phil Blundell
Quoted 19 lines · Martin Jansa
Quoted 17 lines · Khem Raj
Quoted 10 lines · Martin Jansa
Signed-off-by Martin Jansa <[email protected]>--- .../pulseaudio/pulseaudio_0.9.23.bb | 1 + 1 files changed, 1 insertions(+), 0 deletions(-)meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +0
-0meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb do_compile_prepend() { · line 26
Quoted 4 lines
meta/recipes-multimedia/pulseaudio/pulseaudio_0.9.23.bb +1
-0@@ -23,3 +23,4 @@ do_compile_prepend() { cp ${STAGING_LIBDIR}/libltdl* ${S}/libltdl } +ARM_INSTRUCTION_SET = "arm"not averse to patch but it would be nice to know what fails in thumb mode.
Quoted 8 lines
Quoted 1 line
selected processor does not support Thumb mode `swp r1,r2,[r3]'
If it's only that one error (presumably some kind of home-grown atomic operation) then let's just fix the code in question rather than forcing the whole thing to build in ARM-state. Or if the code is hard to fix, patch the makefile so that only the particular source files with the assembly in are built for ARM and let the rest stay as Thumb. Making the whole package be ARM is a bit of a heavyweight fix for that issue.
That's why I've called it work around in cover, because I think that fixing it in old 0.9.23 is not worth it, we can try to fix it in 1.1 when we have it.
And it's not the only one error, I've updated commit message with bigger chunk of log in
http://git.openembedded.org/openembedded-core-contrib/log/?h=jansa/pull
and for om-gta02 (armv4t) it's about
selected processor does not support Thumb mode `mla r3,r7,r0,r3'
while swp was from spitz (armv5te).
Regards,
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
Re: [PATCH 3/8] pulseaudio-0.9.23: force ARM mode
Quoted 2 lines · Martin Jansa
That's why I've called it work around in cover, because I think that fixing it in old 0.9.23 is not worth it, we can try to fix it in 1.1 when we have it.
Fair enough, though as a general rule I don't think the cover letter is the right place for this sort of information. Each patch that you send ought really to be able to stand on its own merits and whatever explanation is necessary ought to be in its own checkin message.
In particular, the cover letter does not form part of the permanent record in the git history, and it is liable to get lost if your patch is incorporated into any kind of consolidated pull.
p.
PATCH 4/8 webkit-gtk: force arm mode to work around binutils segfault
In reply to Martin Jansa--- meta/recipes-sato/webkit/webkit-gtk_svn.bb | 17 +++++++++++++++++ 1 files changed, 17 insertions(+), 0 deletions(-)
meta/recipes-sato/webkit/webkit-gtk_svn.bb +17 -0
@@ -45,6 +45,23 @@ EXTRA_OECONF = "\ EXTRA_AUTORECONF = " -I Source/autotools " + +#| ./Source/JavaScriptCore/heap/HandleTypes.h: In static member function 'static T* JSC::HandleTypes<T>::getFromSlot(JSC::HandleSlot) [with T = JSC::Structure, JSC::HandleTypes<T>::ExternalType = JSC::Structure*, JSC::HandleSlot = JSC::JSValue*]': +#| ./Source/JavaScriptCore/heap/Handle.h:141:79: instantiated from 'JSC::Handle<T>::ExternalType JSC::Handle<T>::get() const [with T = JSC::Structure, JSC::Handle<T>::ExternalType = JSC::Structure*]' +#| ./Source/JavaScriptCore/runtime/ScopeChain.h:39:75: instantiated from here +#| ./Source/JavaScriptCore/heap/HandleTypes.h:38:130: warning: cast from 'JSC::JSCell*' to 'JSC::HandleTypes<JSC::Structure>::ExternalType {aka JSC::Structure*}' increases required alignment of target type [-Wcast-align] +#| {standard input}: Assembler messages: +#| {standard input}:28873: Error: invalid immediate: 983040 is out of range +#| {standard input}:28873: Error: value of 983040 too large for field of 2 bytes at 15110 +#| /OE/shr-core/tmp/sysroots/x86_64-linux/usr/libexec/armv4t-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/4.6.2/as: BFD (GNU Binutils) 2.21.1 assertion fail /OE/shr-core/tmp/work/armv4t-oe-linux-gnueabi/binutils-cross-2.21.1a-r0/binutils-2.21.1/bfd/elf.c:2819 +#| arm-oe-linux-gnueabi-g++: internal compiler error: Segmentation fault (program as) +#| Please submit a full bug report, +#| with preprocessed source if appropriate. +#| See <http://gcc.gnu.org/bugs.html> for instructions. +#| make[1]: *** [Source/JavaScriptCore/jit/libjavascriptcoregtk_1_0_la-JIT.lo] Error 1 +#| make[1]: Leaving directory `/OE/shr-core/tmp/work/armv4t-oe-linux-gnueabi/webkit-gtk-1.5.1+svnr90727-r0' +ARM_INSTRUCTION_SET = "arm" + CONFIGUREOPT_DEPTRACK = "" do_configure_append() {
Signature · 1 line
1.7.7
Re: [PATCH 4/8] webkit-gtk: force arm mode to work around binutils segfault
Quoted 13 lines · Martin Jansa
Signed-off-by Martin Jansa <[email protected]>---
meta/recipes-sato/webkit/webkit-gtk_svn.bb | 17 +++++++++++++++++1 files changed, 17 insertions(+), 0 deletions(-)meta/recipes-sato/webkit/webkit-gtk_svn.bb +0
-0@@ -45,6 +45,23 @@ EXTRA_OECONF = "\EXTRA_AUTORECONF = " -I Source/autotools "
meta/recipes-sato/webkit/webkit-gtk_svn.bb EXTRA_OECONF = "\ · line 57
Quoted 10 lines
meta/recipes-sato/webkit/webkit-gtk_svn.bb +10
-0+ +#| ./Source/JavaScriptCore/heap/HandleTypes.h: In static member function 'static T* JSC::HandleTypes<T>::getFromSlot(JSC::HandleSlot) [with T = JSC::Structure, JSC::HandleTypes<T>::ExternalType = JSC::Structure*, JSC::HandleSlot = JSC::JSValue*]': +#| ./Source/JavaScriptCore/heap/Handle.h:141:79: instantiated from 'JSC::Handle<T>::ExternalType JSC::Handle<T>::get() const [with T = JSC::Structure, JSC::Handle<T>::ExternalType = JSC::Structure*]' +#| ./Source/JavaScriptCore/runtime/ScopeChain.h:39:75: instantiated from here +#| ./Source/JavaScriptCore/heap/HandleTypes.h:38:130: warning: cast from 'JSC::JSCell*' to 'JSC::HandleTypes<JSC::Structure>::ExternalType {aka JSC::Structure*}' increases required alignment of target type [-Wcast-align] +#| {standard input}: Assembler messages: +#| {standard input}:28873: Error: invalid immediate: 983040 is out of range +#| {standard input}:28873: Error: value of 983040 too large for field of 2 bytes at 15110 +#| /OE/shr-core/tmp/sysroots/x86_64-linux/usr/libexec/armv4t-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/4.6.2/as: BFD (GNU Binutils) 2.21.1 assertion fail /OE/shr-core/tmp/work/armv4t-oe-linux-gnueabi/binutils-cross-2.21.1a-r0/binutils-2.21.1/bfd/elf.c:2819 +#| arm-oe-linux-gnueabi-g++: internal compiler error: Segmentation fault (program as)
What it seems is that tables are growing pretty large for thumb to handle. Did you try to use -O1 or -Os with thumb mode ?
Quoted 19 lines
meta/recipes-sato/webkit/webkit-gtk_svn.bb +7
-0+#| Please submit a full bug report, +#| with preprocessed source if appropriate. +#| See <http://gcc.gnu.org/bugs.html> for instructions. +#| make[1]: *** [Source/JavaScriptCore/jit/libjavascriptcoregtk_1_0_la-JIT.lo] Error 1 +#| make[1]: Leaving directory `/OE/shr-core/tmp/work/armv4t-oe-linux-gnueabi/webkit-gtk-1.5.1+svnr90727-r0' +ARM_INSTRUCTION_SET = "arm" +CONFIGUREOPT_DEPTRACK = ""
do_configure_append() {
--
1.7.7_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
PATCH 5/8 apr: add native support
In reply to Martin Jansa* needed for native subversion
--- meta/recipes-support/apr/apr_1.4.5.bb | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
meta/recipes-support/apr/apr_1.4.5.bb +2 -0
@@ -6,6 +6,8 @@ LICENSE = "Apache-2.0" LIC_FILES_CHKSUM = "file://LICENSE;md5=0c35ff3c4c83b89d2f076e315caac28b \ file://include/apr_lib.h;endline=17;md5=ee42fa7575dc40580a9e01c1b75fae96" +BBCLASSEXTEND = "native" + PR = "r1" SRC_URI = "${APACHE_MIRROR}/apr/${BPN}-${PV}.tar.bz2 \
Signature · 1 line
1.7.7
PATCH 6/8 neon: add native support
In reply to Martin Jansa* needed for native subversion
--- meta/recipes-support/neon/neon_0.29.5.bb | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-)
meta/recipes-support/neon/neon_0.29.5.bb +3 -0
@@ -5,9 +5,12 @@ LICENSE = "LGPLv2+" LIC_FILES_CHKSUM = "file://src/COPYING.LIB;md5=f30a9716ef3762e3467a2f62bf790f0a \ file://src/ne_utils.h;beginline=1;endline=20;md5=2caca609538eddaa6f6adf120a218037" DEPENDS = "zlib libxml2 expat time gnutls libproxy" +DEPENDS_virtclass-native = "zlib-native libxml2-native expat-native gnutls-native" PR = "r1" +BBCLASSEXTEND = "native" + SRC_URI = "http://www.webdav.org/${BPN}/${BPN}-${PV}.tar.gz \ file://pkgconfig.patch"
Signature · 1 line
1.7.7
PATCH 7/8 apr-util: add native support
In reply to Martin Jansa* needed for native subversion
* do_configure_append_virtclass-native is needed, because apr_builddir points to /OE/shr-core/tmp/sysroots/x86_64-linux/usr/share/build-1
and we're not installing libtool from apr there
--- meta/recipes-support/apr/apr-util_1.3.12.bb | 8 ++++++++ 1 files changed, 8 insertions(+), 0 deletions(-)
meta/recipes-support/apr/apr-util_1.3.12.bb +8 -0
@@ -3,6 +3,8 @@ HOMEPAGE = "http://apr.apache.org/" SECTION = "libs" DEPENDS = "apr expat gdbm" +BBCLASSEXTEND = "native" + LICENSE = "Apache-2.0" LIC_FILES_CHKSUM = "file://LICENSE;md5=519e0a18e03f7c023070568c14b077bb \ file://include/apu_version.h;endline=17;md5=806685a84e71f10c80144c48eb35df42"
@@ -33,6 +35,12 @@ OE_BINCONFIG_EXTRA_MANGLE = " -e 's:location=source:location=installed:'" do_configure_prepend() { cp ${STAGING_DATADIR}/apr/apr_rules.mk ${S}/build/rules.mk } +do_configure_prepend_virtclass-native() { + cp ${STAGING_DATADIR_NATIVE}/apr/apr_rules.mk ${S}/build/rules.mk +} +do_configure_append_virtclass-native() { + sed -i "s#LIBTOOL=\$(SHELL) \$(apr_builddir)#LIBTOOL=\$(SHELL) ${STAGING_BINDIR_NATIVE}#" ${S}/build/rules.mk +} FILES_${PN} += "${libdir}/apr-util-1/apr_dbm_gdbm-1.so" FILES_${PN}-dev += "${libdir}/aprutil.exp ${libdir}/apr-util-1/apr_dbm_gdbm.*"
Signature · 1 line
1.7.7
PATCH 8/8 subversion: add 1.7.0 with native support and negative D_P for now
In reply to Martin Jansa* be aware that checkouts from 1.7.0 are not compatible with older subversion clients (ie when builder populating distro PREMIRROR is using 1.7.0 all builders need to have also 1.7.0)
* and also 1.7.0 client needs to call svn upgrade in checkout first in order to use it (so if PREMIRROR has tarball from 1.6.x it won't work on client using 1.7.0 unless fetcher2 is improved to detect this and call svn upgrade)
* tested on SHR distribution
http://wiki.shr-project.org/trac/wiki/Building%20SHR#subversion1.7inshr-chroot
* only missing part is to add subversion-native dependency, so that native subversion is built, before building ie elementary (because EFL are using svnversion from configure.ac to detect source revision and .svn dir needs to be from compatible version).
* read http://subversion.apache.org/docs/release-notes/1.7.html
--- .../subversion/subversion-1.7.0/libtool2.patch | 15 ++++++++ .../subversion/subversion_1.7.0.bb | 37 ++++++++++++++++++++ 2 files changed, 52 insertions(+), 0 deletions(-) create mode 100644 meta/recipes-devtools/subversion/subversion-1.7.0/libtool2.patch create mode 100644 meta/recipes-devtools/subversion/subversion_1.7.0.bb
meta/recipes-devtools/subversion/subversion-1.7.0/libtool2.patch new file +15 -0
@@ -0,0 +1,15 @@ +Upstream-Status: Inappropriate [embedded specific] + +--- a/configure.ac 2011-10-20 21:56:02.230663987 +0200 ++++ b/configure.ac 2011-08-17 15:01:30.000000000 +0200 +@@ -227,8 +227,8 @@ + LIBTOOL="$sh_libtool" + SVN_LIBTOOL="$sh_libtool" + else +- sh_libtool="$abs_builddir/libtool" +- SVN_LIBTOOL="\$(SHELL) $sh_libtool" ++ sh_libtool="$abs_builddir/$host_alias-libtool" ++ SVN_LIBTOOL="\$(SHELL) \$(abs_builddir)/$host_alias-libtool" + fi + AC_SUBST(SVN_LIBTOOL) +
meta/recipes-devtools/subversion/subversion_1.7.0.bb new file +37 -0
@@ -0,0 +1,37 @@ +DESCRIPTION = "The Subversion (svn) client" +SECTION = "console/network" +DEPENDS = "apr-util neon" +RDEPENDS_${PN} = "neon" +LICENSE = "Apache-2" +HOMEPAGE = "http://subversion.tigris.org" + +BBCLASSEXTEND = "native" + +# negative, because of new checkout format in 1.7.0 +# and distro PREMIRRORs need to be in sync with users +DEFAULT_PREFERENCE = "-1" + +SRC_URI = "${APACHE_MIRROR}/${BPN}/${BPN}-${PV}.tar.bz2 \ + file://libtool2.patch \ +" +SRC_URI[md5sum] = "930e6644a1b6094efd268fde6a318f04" +SRC_URI[sha256sum] = "64fd5f263a80e609717a3ca42f1f2625606a5c4a40a85716f82c866033780978" + +LIC_FILES_CHKSUM = "file://LICENSE;md5=4a14fd2da3134e40a087eb4326a4ecd4" + +EXTRA_OECONF = " \ + --without-berkeley-db --without-apxs --without-apache \ + --without-swig --with-apr=${STAGING_BINDIR_CROSS} \ + --with-apr-util=${STAGING_BINDIR_CROSS}" + +inherit autotools + +export LDFLAGS += " -L${STAGING_LIBDIR} " + +acpaths = "-I build/ -I build/ac-macros/" + +do_configure_prepend () { + rm -f ${S}/libtool + rm -f ${S}/build/libtool.m4 + sed -i -e 's:with_sasl="/usr/local":with_sasl="${STAGING_DIR}":' ${S}/build/ac-macros/sasl.m4 +}
Signature · 1 line
1.7.7
PATCH 1/1 base.bbclass: add subversion-native to DEPENDS if there is svn:// in SRC_URI
--- meta/classes/base.bbclass | 7 +++++++ 1 files changed, 7 insertions(+), 0 deletions(-)
meta/classes/base.bbclass +7 -0
@@ -401,6 +401,13 @@ python () { bb.note("SKIPPING %s because it's %s" % (pn, this_license)) raise bb.parse.SkipPackage("incompatible with license %s" % this_license) + # Svn packages should DEPEND on subversion-native + srcuri = bb.data.getVar('SRC_URI', d, 1) + if "svn://" in srcuri: + depends = bb.data.getVarFlag('do_fetch', 'depends', d) or "" + depends = depends + " subversion-native:do_populate_sysroot" + bb.data.setVarFlag('do_fetch', 'depends', depends, d) + # Git packages should DEPEND on git-native srcuri = bb.data.getVar('SRC_URI', d, 1) if "git://" in srcuri:
Signature · 1 line
1.7.7
Re: [PATCH] base.bbclass: add subversion-native to DEPENDS if there is svn:// in SRC_URI
Quoted 13 lines · Martin Jansa
Signed-off-by Martin Jansa <[email protected]>--- meta/classes/base.bbclass | 7 +++++++ 1 files changed, 7 insertions(+), 0 deletions(-)meta/classes/base.bbclass +0
-0@@ -401,6 +401,13 @@ python () { bb.note("SKIPPING %s because it's %s" % (pn, this_license)) raise bb.parse.SkipPackage("incompatible with license %s" % this_license)
meta/classes/base.bbclass python () { · line 413
Quoted 10 lines
meta/classes/base.bbclass +7
-0+ # Svn packages should DEPEND on subversion-native + srcuri = bb.data.getVar('SRC_URI', d, 1) + if "svn://" in srcuri: + depends = bb.data.getVarFlag('do_fetch', 'depends', d) or "" + depends = depends + " subversion-native:do_populate_sysroot" + bb.data.setVarFlag('do_fetch', 'depends', depends, d) + # Git packages should DEPEND on git-native srcuri = bb.data.getVar('SRC_URI', d, 1) if "git://" in srcuri:
For this to work correctly, we need to add BBCLASSEXTEND = "native" also to subversion_1.6* otherwise everybody will get subversion-native-1.7.0
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
PATCH 1/1 subversion-1.6.15: add native support too
--- .../subversion/subversion_1.6.15.bb | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
meta/recipes-devtools/subversion/subversion_1.6.15.bb +2 -0
@@ -5,6 +5,8 @@ RDEPENDS_${PN} = "neon" LICENSE = "Apache-2" HOMEPAGE = "http://subversion.tigris.org" +BBCLASSEXTEND = "native" + PR = "r1" SRC_URI = "http://subversion.tigris.org/downloads/${BPN}-${PV}.tar.bz2 \
Signature · 1 line
1.7.7
Re: [PATCH] subversion-1.6.15: add native support too
Quoted 10 lines · Martin Jansa
Signed-off-by Martin Jansa<[email protected]>--- .../subversion/subversion_1.6.15.bb | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)meta/recipes-devtools/subversion/subversion_1.6.15.bb +0
-0
meta/recipes-devtools/subversion/subversion_1.6.15.bb RDEPENDS_${PN} = "neon" · line 12
Quoted 8 lines
meta/recipes-devtools/subversion/subversion_1.6.15.bb +2
-0@@ -5,6 +5,8 @@ RDEPENDS_${PN} = "neon" LICENSE = "Apache-2" HOMEPAGE = "http://subversion.tigris.org" +BBCLASSEXTEND = "native" + PR = "r1" SRC_URI = "http://subversion.tigris.org/downloads/${BPN}-${PV}.tar.bz2 \
Pulled this one also into OE-Core
Thanks
Sau!
Re: [PATCH] base.bbclass: add subversion-native to DEPENDS if there is svn:// in SRC_URI
In reply to Martin JansaQuoted 13 lines · Martin Jansa
Signed-off-by Martin Jansa<[email protected]>--- meta/classes/base.bbclass | 7 +++++++ 1 files changed, 7 insertions(+), 0 deletions(-)meta/classes/base.bbclass +0
-0@@ -401,6 +401,13 @@ python () { bb.note("SKIPPING %s because it's %s" % (pn, this_license)) raise bb.parse.SkipPackage("incompatible with license %s" % this_license)
meta/classes/base.bbclass python () { · line 413
Quoted 10 lines
meta/classes/base.bbclass +7
-0+ # Svn packages should DEPEND on subversion-native + srcuri = bb.data.getVar('SRC_URI', d, 1) + if "svn://" in srcuri: + depends = bb.data.getVarFlag('do_fetch', 'depends', d) or "" + depends = depends + " subversion-native:do_populate_sysroot" + bb.data.setVarFlag('do_fetch', 'depends', depends, d) + # Git packages should DEPEND on git-native srcuri = bb.data.getVar('SRC_URI', d, 1) if "git://" in srcuri:
This change introduces a circular dependency when I tried to build.
Sau!
Quoted 20 lines · Martin Jansa
Hi,
arm mode was forced without PR bump, because if it was failing for someone he will notice.
Be carefull with using subversion-1.7.0 (read the commit message).
Pulseaudio-1.1 was released, so hopefully this work arounds won't be needed there or they could be resolved better with upgrade.
Regards,
The following changes since commit
99da9a4e65f9dffb04efc3ad60125194c476d6b3:distro-tracking-fields: update fields for tzdata and gst-plugins-good (2011-10-20 13:07:16 +0100)are available in the git repository at:
git://git.openembedded.org/openembedded-core-contribjansa/pull2
http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=jansa/pull2
Quoted 11 lines
Martin Jansa (8):
pulseaudio-0.9.23: inherit perlnative to work around build on host
without XML/Parser.pm
libatomics-ops: force ARM mode
pulseaudio-0.9.23: force ARM mode
webkit-gtk: force arm mode to work around binutils segfault
apr: add native support
neon: add native support
apr-util: add native support
subversion: add 1.7.0 with native support and negativeD_Pfor now
Of the above patches, the pulse audio and adding native support for apr, neon and apr-util were taken, the other are on hold pending different solutions.
Thanks
Sau!
Quoted 12 lines
.../subversion/subversion-1.7.0/libtool2.patch | 15 ++++++++ .../subversion/subversion_1.7.0.bb | 37 ++++++++++++++++++++ .../pulseaudio/libatomics-ops_1.2.bb | 2 + .../pulseaudio/pulseaudio_0.9.23.bb | 3 +- meta/recipes-sato/webkit/webkit-gtk_svn.bb | 17 +++++++++ meta/recipes-support/apr/apr-util_1.3.12.bb | 8 ++++ meta/recipes-support/apr/apr_1.4.5.bb | 2 + meta/recipes-support/neon/neon_0.29.5.bb | 3 ++ 8 files changed, 86 insertions(+), 1 deletions(-)create mode 100644
meta/recipes-devtools/subversion/subversion-1.7.0/libtool2.patchcreate mode 100644 meta/recipes-devtools/subversion/subversion_1.7.0.bb
Quoted 32 lines · Saul Wold
Quoted 20 lines · Martin Jansa
Hi,
arm mode was forced without PR bump, because if it was failing for someone he will notice.
Be carefull with using subversion-1.7.0 (read the commit message).
Pulseaudio-1.1 was released, so hopefully this work arounds won't be needed there or they could be resolved better with upgrade.
Regards,
The following changes since commit
99da9a4e65f9dffb04efc3ad60125194c476d6b3:distro-tracking-fields: update fields for tzdata and gst-plugins-good (2011-10-20 13:07:16 +0100)are available in the git repository at:
git://git.openembedded.org/openembedded-core-contribjansa/pull2
http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=jansa/pull2Quoted 11 lines
Martin Jansa (8):
pulseaudio-0.9.23: inherit perlnative to work around build on host
without XML/Parser.pm
libatomics-ops: force ARM mode
pulseaudio-0.9.23: force ARM mode
webkit-gtk: force arm mode to work around binutils segfault
apr: add native support
neon: add native support
apr-util: add native support
subversion: add 1.7.0 with native support and negativeD_Pfor now
Quoted 3 lines
Of the above patches, the pulse audio and adding native support for apr, neon and apr-util were taken, the other are on hold pending different solutions.
I've added error log to pulseaudio commit and consolidated all my pending patches to single jansa/pull branch, but different (better) solutions are welcome :).
Cheers,
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
Quoted 20 lines · Martin Jansa
Hi,
arm mode was forced without PR bump, because if it was failing for someone he will notice.
Be carefull with using subversion-1.7.0 (read the commit message).
Pulseaudio-1.1 was released, so hopefully this work arounds won't be needed there or they could be resolved better with upgrade.
Regards,
The following changes since commit
99da9a4e65f9dffb04efc3ad60125194c476d6b3:distro-tracking-fields: update fields for tzdata and gst-plugins-good (2011-10-20 13:07:16 +0100)are available in the git repository at:
git://git.openembedded.org/openembedded-core-contribjansa/pull2
http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=jansa/pull2
Quoted 6 lines
Martin Jansa (8):
pulseaudio-0.9.23: inherit perlnative to work around build on host
without XML/Parser.pm
libatomics-ops: force ARM mode
pulseaudio-0.9.23: force ARM mode
webkit-gtk: force arm mode to work around binutils segfault
Martin,
What is the status of these various workarounds, did you do any further work to determine root causes, or do we still need these patches to do in?
Thanks
Sau!
Quoted 17 lines
apr: add native support
neon: add native support
apr-util: add native support
subversion: add 1.7.0 with native support and negativeD_Pfor now.../subversion/subversion-1.7.0/libtool2.patch | 15 ++++++++ .../subversion/subversion_1.7.0.bb | 37 ++++++++++++++++++++ .../pulseaudio/libatomics-ops_1.2.bb | 2 + .../pulseaudio/pulseaudio_0.9.23.bb | 3 +- meta/recipes-sato/webkit/webkit-gtk_svn.bb | 17 +++++++++ meta/recipes-support/apr/apr-util_1.3.12.bb | 8 ++++ meta/recipes-support/apr/apr_1.4.5.bb | 2 + meta/recipes-support/neon/neon_0.29.5.bb | 3 ++ 8 files changed, 86 insertions(+), 1 deletions(-)create mode 100644
meta/recipes-devtools/subversion/subversion-1.7.0/libtool2.patchcreate mode 100644 meta/recipes-devtools/subversion/subversion_1.7.0.bb
Quoted 27 lines · Saul Wold
Quoted 20 lines · Martin Jansa
Hi,
arm mode was forced without PR bump, because if it was failing for someone he will notice.
Be carefull with using subversion-1.7.0 (read the commit message).
Pulseaudio-1.1 was released, so hopefully this work arounds won't be needed there or they could be resolved better with upgrade.
Regards,
The following changes since commit
99da9a4e65f9dffb04efc3ad60125194c476d6b3:distro-tracking-fields: update fields for tzdata and gst-plugins-good (2011-10-20 13:07:16 +0100)are available in the git repository at:
git://git.openembedded.org/openembedded-core-contribjansa/pull2
http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=jansa/pull2Quoted 6 lines
Martin Jansa (8):
pulseaudio-0.9.23: inherit perlnative to work around build on host
without XML/Parser.pm
libatomics-ops: force ARM mode
pulseaudio-0.9.23: force ARM mode
webkit-gtk: force arm mode to work around binutils segfault
Quoted 4 lines
Martin,
What is the status of these various workarounds, did you do any further work to determine root causes, or do we still need these patches to do in?
I've only updated commit messages in branch and added the same workaround for aspell which also fails for thumb.
All in jansa/pull
Quoted 21 lines
Thanks
Sau!Quoted 17 lines
apr: add native support
neon: add native support
apr-util: add native support
subversion: add 1.7.0 with native support and negativeD_Pfor now.../subversion/subversion-1.7.0/libtool2.patch | 15 ++++++++ .../subversion/subversion_1.7.0.bb | 37 ++++++++++++++++++++ .../pulseaudio/libatomics-ops_1.2.bb | 2 + .../pulseaudio/pulseaudio_0.9.23.bb | 3 +- meta/recipes-sato/webkit/webkit-gtk_svn.bb | 17 +++++++++ meta/recipes-support/apr/apr-util_1.3.12.bb | 8 ++++ meta/recipes-support/apr/apr_1.4.5.bb | 2 + meta/recipes-support/neon/neon_0.29.5.bb | 3 ++ 8 files changed, 86 insertions(+), 1 deletions(-)create mode 100644
meta/recipes-devtools/subversion/subversion-1.7.0/libtool2.patchcreate mode 100644 meta/recipes-devtools/subversion/subversion_1.7.0.bb
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]