[PATCH 4/7] xorg-[lib-common|proto-common]: disable use of fop document generation

[PATCH 4/7] xorg-[lib-common|proto-common]: disable use of fop document generation

[Joshua Lock]

From: Joshua Lock <josh@linux.intel.com>

On Fedora 15 I see a huge Java backtrace when document generation runs for
some xorg libs. As fop is automatically detected, with the possibility of
detecting fop on the host whilst doing target builds, the safest bet is to
explicitly disable fop for document generation.

Signed-off-by: Joshua Lock <josh@linux.intel.com>
---
 meta/recipes-graphics/xorg-lib/xorg-lib-common.inc |    2 +-
 .../xorg-proto/xorg-proto-common.inc               |    1 +
 2 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/meta/recipes-graphics/xorg-lib/xorg-lib-common.inc b/meta/recipes-graphics/xorg-lib/xorg-lib-common.inc
index 4ed8c86..d843d65 100644
--- a/meta/recipes-graphics/xorg-lib/xorg-lib-common.inc
+++ b/meta/recipes-graphics/xorg-lib/xorg-lib-common.inc
@@ -13,4 +13,4 @@ S = "${WORKDIR}/${XORG_PN}-${PV}"
 
 inherit autotools pkgconfig
 
-EXTRA_OECONF = "--enable-malloc0returnsnull"
+EXTRA_OECONF = "--enable-malloc0returnsnull --with-fop=no"
diff --git a/meta/recipes-graphics/xorg-proto/xorg-proto-common.inc b/meta/recipes-graphics/xorg-proto/xorg-proto-common.inc
index 92aa1fe..ab1af28 100644
--- a/meta/recipes-graphics/xorg-proto/xorg-proto-common.inc
+++ b/meta/recipes-graphics/xorg-proto/xorg-proto-common.inc
@@ -14,6 +14,7 @@ S = "${WORKDIR}/${XORG_PN}-${PV}"
 DEPENDS = "util-macros"
 
 inherit autotools pkgconfig
+EXTRA_OECONF = "--with-fop=no"
 
 # ${PN} is empty so we need to tweak -dev and -dbg package dependencies
 RDEPENDS_${PN}-dev = ""
-- 
1.7.4.1

[PATCH 5/7] libx11: disable building of specs

[Joshua Lock]

From: Joshua Lock <josh@linux.intel.com>

Generating Postscript specs fails on Fedora 15, I don't *think* we need them
so disable them.

Signed-off-by: Joshua Lock <josh@linux.intel.com>
---
 meta/recipes-graphics/xorg-lib/libx11.inc |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/meta/recipes-graphics/xorg-lib/libx11.inc b/meta/recipes-graphics/xorg-lib/libx11.inc
index 267a03a..030b0db 100644
--- a/meta/recipes-graphics/xorg-lib/libx11.inc
+++ b/meta/recipes-graphics/xorg-lib/libx11.inc
@@ -11,7 +11,7 @@ PROVIDES = "virtual/libx11"
 XORG_PN = "libX11"
 LEAD_SONAME = "libX11.so"
 
-EXTRA_OECONF += "--with-keysymdef=${STAGING_INCDIR}/X11/keysymdef.h"
+EXTRA_OECONF += "--with-keysymdef=${STAGING_INCDIR}/X11/keysymdef.h --disable-specs"
 
 FILES_${PN} += "${datadir}/X11/XKeysymDB ${datadir}/X11/XErrorDB ${libdir}/X11/Xcms.txt"
 FILES_${PN}-locale += "${datadir}/X11/locale ${libdir}/X11/locale"
-- 
1.7.4.1

[PATCH 6/7] openjade: fix build with GCC 4.6

[Joshua Lock]

From: Joshua Lock <josh@linux.intel.com>

In GCC 4.6 the compiler no longer allows objects of const-qualified type to
be default initialized unless the type has a user-declared default
constructor.

Patch from Gentoo bugzilla: http://bugs.gentoo.org/show_bug.cgi?id=358021

Signed-off-by: Joshua Lock <josh@linux.intel.com>
---
 .../user-declared-default-constructor.patch        |   88 ++++++++++++++++++++
 .../openjade/openjade-native_1.3.2.bb              |    5 +-
 2 files changed, 91 insertions(+), 2 deletions(-)
 create mode 100644 meta/recipes-devtools/openjade/openjade-1.3.2/user-declared-default-constructor.patch

diff --git a/meta/recipes-devtools/openjade/openjade-1.3.2/user-declared-default-constructor.patch b/meta/recipes-devtools/openjade/openjade-1.3.2/user-declared-default-constructor.patch
new file mode 100644
index 0000000..2f2adfe
--- /dev/null
+++ b/meta/recipes-devtools/openjade/openjade-1.3.2/user-declared-default-constructor.patch
@@ -0,0 +1,88 @@
+In GCC 4.6 the compiler no longer allows objects of const-qualified type to
+be default initialized unless the type has a user-declared default
+constructor.
+
+Patch from Gentoo bugzilla: http://bugs.gentoo.org/show_bug.cgi?id=358021
+
+Gentoo Bugzilla description follows:
+"If a class or struct has no user-defined default constructor, C++ doesn't
+allow you to default construct a const instance of it.
+
+https://bugs.gentoo.org/358021
+http://clang.llvm.org/compatibility.html#default_init_const
+http://gcc.gnu.org/PR44499"
+
+--- a/jade/TeXFOTBuilder.cxx
++++ b/jade/TeXFOTBuilder.cxx
+@@ -88,6 +88,8 @@ public:
+       value.convertString(nic_.placement);
+       }
+     ExtensionFlowObj *copy() const { return new PageFloatFlowObj(*this); }
++  public:
++    PageFloatFlowObj() {}
+   private:
+     PageFloatNIC nic_;
+     StringC name_;
+@@ -101,6 +103,8 @@ public:
+       fotb.endPageFootnote();
+     }
+     ExtensionFlowObj *copy() const { return new PageFootnoteFlowObj(*this); }
++  public:
++    PageFootnoteFlowObj() {}
+   private:
+   };
+   //////////////////////////////////////////////////////////////////////
+--- a/jade/TransformFOTBuilder.cxx
++++ b/jade/TransformFOTBuilder.cxx
+@@ -41,6 +41,7 @@ public:
+   };
+   class EntityRefFlowObj : public TransformExtensionFlowObj {
+   public:
++    EntityRefFlowObj() {}
+     void atomic(TransformFOTBuilder &fotb, const NodePtr &) const {
+       fotb.entityRef(name_);
+     }
+@@ -56,6 +57,7 @@ public:
+   };
+   class ProcessingInstructionFlowObj : public TransformExtensionFlowObj {
+   public:
++    ProcessingInstructionFlowObj() {}
+     void atomic(TransformFOTBuilder &fotb, const NodePtr &) const {
+       fotb.processingInstruction(data_);
+     }
+@@ -98,6 +100,8 @@ public:
+       }
+     }
+     ExtensionFlowObj *copy() const { return new EmptyElementFlowObj(*this); }
++  public:
++    EmptyElementFlowObj() {}
+   private:
+     ElementNIC nic_;
+   };
+@@ -133,6 +137,8 @@ public:
+       }
+     }
+     ExtensionFlowObj *copy() const { return new ElementFlowObj(*this); }
++  public:
++    ElementFlowObj() {}
+   private:
+     ElementNIC nic_;
+   };
+@@ -150,6 +156,8 @@ public:
+       value.convertString(systemId_);
+     }
+     ExtensionFlowObj *copy() const { return new EntityFlowObj(*this); }
++  public:
++    EntityFlowObj() {}
+   private:
+     StringC systemId_;
+   };
+@@ -174,6 +182,8 @@ public:
+       }
+     }
+     ExtensionFlowObj *copy() const { return new DocumentTypeFlowObj(*this); }
++  public:
++    DocumentTypeFlowObj() {}
+   private:
+     DocumentTypeNIC nic_;
+   };
diff --git a/meta/recipes-devtools/openjade/openjade-native_1.3.2.bb b/meta/recipes-devtools/openjade/openjade-native_1.3.2.bb
index 0de8b96..5c1037a 100644
--- a/meta/recipes-devtools/openjade/openjade-native_1.3.2.bb
+++ b/meta/recipes-devtools/openjade/openjade-native_1.3.2.bb
@@ -7,13 +7,14 @@ SECTION = "base"
 LICENSE = "BSD"
 LIC_FILES_CHKSUM = "file://COPYING;md5=641ff1e4511f0a87044ad42f87cb1045"
 
-PR = "r1"
+PR = "r2"
 
 DEPENDS = "opensp-native sgml-common-native"
 RDEPENDS_${PN} = "sgml-common"
 
 SRC_URI = "${SOURCEFORGE_MIRROR}/openjade/openjade-${PV}.tar.gz \
-           file://makefile.patch"
+           file://makefile.patch \
+	   file://user-declared-default-constructor.patch"
 
 SRC_URI[md5sum] = "7df692e3186109cc00db6825b777201e"
 SRC_URI[sha256sum] = "1d2d7996cc94f9b87d0c51cf0e028070ac177c4123ecbfd7ac1cb8d0b7d322d1"
-- 
1.7.4.1

[PATCH 7/7] xserver-xf86: explicitly disable fop document generation

[Joshua Lock]

From: Joshua Lock <josh@linux.intel.com>

Signed-off-by: Joshua Lock <josh@linux.intel.com>
---
 .../xorg-xserver/xserver-xf86-common.inc           |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/meta/recipes-graphics/xorg-xserver/xserver-xf86-common.inc b/meta/recipes-graphics/xorg-xserver/xserver-xf86-common.inc
index f9ea0e4..b1a0f08 100644
--- a/meta/recipes-graphics/xorg-xserver/xserver-xf86-common.inc
+++ b/meta/recipes-graphics/xorg-xserver/xserver-xf86-common.inc
@@ -34,3 +34,4 @@ FILES_${PN}-dbg += "${libdir}/xorg/modules/.debug \
 
 SRC_URI += "file://macro_tweak.patch"
 
+EXTRA_OECONF = "--with-fop=no"
-- 
1.7.4.1

[PATCH 0/7] Patch series against Bernard for Fedora 15 users

[Joshua Lock]

From: Joshua Lock <josh@linux.intel.com>

I decided to punish myself by testing how well our branch fares on Fedora 15.

I don't think any of these should be too controversial and with the series
applied I am able to build both minimal and sato images on my Fedora 15 box.

Pull URL: git://git.pokylinux.org/poky-contrib.git
  Branch: josh/bernard-f15
  Browse: http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=josh/bernard-f15

Thanks,
    Joshua Lock <josh@linux.intel.com>
---


Joshua Lock (7):
  elfutils: fix builds with gcc 4.6
  python-native: add missing SRC_URI hashes
  libx11: add missing SRC_URI hashes
  xorg-[lib-common|proto-common]: disable use of fop document
    generation
  libx11: disable building of specs
  openjade: fix build with GCC 4.6
  xserver-xf86: explicitly disable fop document generation

 meta/recipes-devtools/elfutils/elfutils_0.148.bb   |    4 +
 .../user-declared-default-constructor.patch        |   88 ++++++++++++++++++++
 .../openjade/openjade-native_1.3.2.bb              |    5 +-
 .../recipes-devtools/python/python-native_2.6.6.bb |    3 +
 meta/recipes-graphics/xorg-lib/libx11.inc          |    2 +-
 meta/recipes-graphics/xorg-lib/libx11_1.3.4.bb     |    3 +
 meta/recipes-graphics/xorg-lib/xorg-lib-common.inc |    2 +-
 .../xorg-proto/xorg-proto-common.inc               |    1 +
 .../xorg-xserver/xserver-xf86-common.inc           |    1 +
 9 files changed, 105 insertions(+), 4 deletions(-)
 create mode 100644 meta/recipes-devtools/openjade/openjade-1.3.2/user-declared-default-constructor.patch

-- 
1.7.4.1

[PATCH 1/7] elfutils: fix builds with gcc 4.6

[Joshua Lock]

From: Joshua Lock <josh@linux.intel.com>

gcc 4.6 (as used in Fedora 15) adds some extra warnings which are
included with Werror. The new unused-but-set variable warning causes
an error in libasm of elfutils. Work around this by removing
unused-but-set from Werror.

Signed-off-by: Joshua Lock <josh@linux.intel.com>
---
 meta/recipes-devtools/elfutils/elfutils_0.148.bb |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/meta/recipes-devtools/elfutils/elfutils_0.148.bb b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
index b2f700e..c395be8 100644
--- a/meta/recipes-devtools/elfutils/elfutils_0.148.bb
+++ b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
@@ -40,6 +40,10 @@ SRC_URI += "\
 
 inherit autotools
 
+# GCC 4.6.0 raises an unused-but-set warning in libasm, for now remove
+# this warning from Werror
+CFLAGS_virtclass-native += "-Wno-error=unused-but-set-variable"
+
 EXTRA_OECONF = "--program-prefix=eu-"
 
 do_configure_prepend() {
-- 
1.7.4.1

[PATCH 2/7] python-native: add missing SRC_URI hashes

[Joshua Lock]

From: Joshua Lock <josh@linux.intel.com>

Signed-off-by: Joshua Lock <josh@linux.intel.com>
---
 .../recipes-devtools/python/python-native_2.6.6.bb |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/meta/recipes-devtools/python/python-native_2.6.6.bb b/meta/recipes-devtools/python/python-native_2.6.6.bb
index 2ad10c9..11956a4 100644
--- a/meta/recipes-devtools/python/python-native_2.6.6.bb
+++ b/meta/recipes-devtools/python/python-native_2.6.6.bb
@@ -14,6 +14,9 @@ SRC_URI = "http://www.python.org/ftp/python/${PV}/Python-${PV}.tar.bz2 \
            file://nohostlibs.patch"
 S = "${WORKDIR}/Python-${PV}"
 
+SRC_URI[md5sum] = "cf4e6881bb84a7ce6089e4a307f71f14"
+SRC_URI[sha256sum] = "134c5e0736bae2e5570d0b915693374f11108ded63c35a23a35d282737d2ce83"
+
 inherit native
 
 EXTRA_OEMAKE = '\
-- 
1.7.4.1

[PATCH 3/7] libx11: add missing SRC_URI hashes

[Joshua Lock]

From: Joshua Lock <josh@linux.intel.com>

Signed-off-by: Joshua Lock <josh@linux.intel.com>
---
 meta/recipes-graphics/xorg-lib/libx11_1.3.4.bb |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/meta/recipes-graphics/xorg-lib/libx11_1.3.4.bb b/meta/recipes-graphics/xorg-lib/libx11_1.3.4.bb
index aaeb38e..c572d5e 100644
--- a/meta/recipes-graphics/xorg-lib/libx11_1.3.4.bb
+++ b/meta/recipes-graphics/xorg-lib/libx11_1.3.4.bb
@@ -10,6 +10,9 @@ SRC_URI += "file://x11_disable_makekeys.patch \
             file://nodolt.patch \
             file://include_fix.patch"
 
+SRC_URI[md5sum] = "f65c9c7ecbfb64c19dbd7927160d63fd"
+SRC_URI[sha256sum] = "88d7238ce5f7cd123450567de7a3b56a43556e4ccc45df38b8324147c889a844"
+
 DEPENDS += "bigreqsproto xproto xextproto xtrans libxau xcmiscproto \
             libxdmcp xf86bigfontproto kbproto inputproto xproto-native gettext"
 
-- 
1.7.4.1

Re: [PATCH 4/7] xorg-[lib-common|proto-common]: disable use of fop document generation

[Colin Walters]

On Fri, Apr 1, 2011 at 9:38 AM, Joshua Lock <josh@linux.intel.com> wrote:
> From: Joshua Lock <josh@linux.intel.com>
>
> On Fedora 15 I see a huge Java backtrace when document generation runs for
> some xorg libs. As fop is automatically detected, with the possibility of
> detecting fop on the host whilst doing target builds, the safest bet is to
> explicitly disable fop for document generation.

This is actually a bug in the GCJ stack; it should go away if you use
OpenJDK.  Try using "update-alternatives --config java".

But yeah, I doubt any embeds care about the docs, and certainly fop
pulls in a frightening dependency chain; so cool by me to disable.

Re: [PATCH 4/7] xorg-[lib-common|proto-common]: disable use of fop document generation

[Martin Jansa]

On Fri, Apr 1, 2011 at 3:38 PM, Joshua Lock <josh@linux.intel.com> wrote:
> From: Joshua Lock <josh@linux.intel.com>
>
> On Fedora 15 I see a huge Java backtrace when document generation runs for
> some xorg libs. As fop is automatically detected, with the possibility of
> detecting fop on the host whilst doing target builds, the safest bet is to
> explicitly disable fop for document generation.
>
> Signed-off-by: Joshua Lock <josh@linux.intel.com>
> ---
>  meta/recipes-graphics/xorg-lib/xorg-lib-common.inc |    2 +-
>  .../xorg-proto/xorg-proto-common.inc               |    1 +
>  2 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/meta/recipes-graphics/xorg-lib/xorg-lib-common.inc b/meta/recipes-graphics/xorg-lib/xorg-lib-common.inc
> index 4ed8c86..d843d65 100644
> --- a/meta/recipes-graphics/xorg-lib/xorg-lib-common.inc
> +++ b/meta/recipes-graphics/xorg-lib/xorg-lib-common.inc
> @@ -13,4 +13,4 @@ S = "${WORKDIR}/${XORG_PN}-${PV}"
>
>  inherit autotools pkgconfig
>
> -EXTRA_OECONF = "--enable-malloc0returnsnull"
> +EXTRA_OECONF = "--enable-malloc0returnsnull --with-fop=no"
> diff --git a/meta/recipes-graphics/xorg-proto/xorg-proto-common.inc b/meta/recipes-graphics/xorg-proto/xorg-proto-common.inc
> index 92aa1fe..ab1af28 100644
> --- a/meta/recipes-graphics/xorg-proto/xorg-proto-common.inc
> +++ b/meta/recipes-graphics/xorg-proto/xorg-proto-common.inc
> @@ -14,6 +14,7 @@ S = "${WORKDIR}/${XORG_PN}-${PV}"
>  DEPENDS = "util-macros"
>
>  inherit autotools pkgconfig
> +EXTRA_OECONF = "--with-fop=no"

after this I get

| checking for xmlto... no
| configure: WARNING: xmlto not found - documentation targets will be skipped
| configure: error: --with-fop expects 'yes' or 'no'
| FATAL: oe_runconf failed
| ERROR: Function 'do_configure' failed (see
/OE/shr-core/tmp/work/x86_64-linux/xproto-native-1_7.0.20-r0/temp/log.do_configure.29076
for further information)
NOTE: package xproto-native-1_7.0.20-r0: task do_configure: Failed
ERROR: Task 957
(virtual:native:/OE/shr-core/openembedded-core/meta/recipes-graphics/xorg-proto/xproto_7.0.20.bb,
do_configure) failed with exit code '1'

because there is
--with-fop=no--enable-specs=no

so you had to put space after no, or before this

OE nokia900@shr ~/shr-core $ grep OECONF
openembedded-core/meta/recipes-graphics/xorg-proto/xproto_7.0.20.bb
EXTRA_OECONF_append = "--enable-specs=no"

>
>  # ${PN} is empty so we need to tweak -dev and -dbg package dependencies
>  RDEPENDS_${PN}-dev = ""
> --
> 1.7.4.1
>
> _______________________________________________
> poky mailing list
> poky@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/poky
>

Re: [PATCH 1/7] elfutils: fix builds with gcc 4.6

[Martin Jansa]

On Fri, Apr 1, 2011 at 3:40 PM, Joshua Lock <josh@linux.intel.com> wrote:
> From: Joshua Lock <josh@linux.intel.com>
>
> gcc 4.6 (as used in Fedora 15) adds some extra warnings which are
> included with Werror. The new unused-but-set variable warning causes
> an error in libasm of elfutils. Work around this by removing
> unused-but-set from Werror.
>
> Signed-off-by: Joshua Lock <josh@linux.intel.com>
> ---
>  meta/recipes-devtools/elfutils/elfutils_0.148.bb |    4 ++++
>  1 files changed, 4 insertions(+), 0 deletions(-)
>
> diff --git a/meta/recipes-devtools/elfutils/elfutils_0.148.bb b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> index b2f700e..c395be8 100644
> --- a/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> +++ b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> @@ -40,6 +40,10 @@ SRC_URI += "\
>
>  inherit autotools
>
> +# GCC 4.6.0 raises an unused-but-set warning in libasm, for now remove
> +# this warning from Werror
> +CFLAGS_virtclass-native += "-Wno-error=unused-but-set-variable"
> +
>  EXTRA_OECONF = "--program-prefix=eu-"

this breaks all other distros not yet using gcc-4.6..


configure:2892: checking whether the C compiler works
configure:2914: gcc -Wno-error=unused-but-set-variable
-isystem/OE/shr-core/tmp/sysroots/x86_64-linux/usr/include
-L/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
-Wl,-rpath-link,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
-Wl,-rpath,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib -Wl,-O1
conftest.c  >&5
cc1: error: -Werror=unused-but-set-variable: No option -Wunused-but-set-variable
configure:2918: $? = 1
configure:2956: result: no
configure: failed program was:
| /* confdefs.h */
| #define PACKAGE_NAME "Red Hat elfutils"
| #define PACKAGE_TARNAME "elfutils"
| #define PACKAGE_VERSION "0.148"
| #define PACKAGE_STRING "Red Hat elfutils 0.148"
| #define PACKAGE_BUGREPORT "http://bugzilla.redhat.com/bugzilla/"
| #define PACKAGE_URL ""
| #define PACKAGE "elfutils"
| #define VERSION "0.148"
| #define MODVERSION "Build on jama 2011-04-04T23:34:25+0200"
| /* end confdefs.h.  */
|
| int
| main ()
| {
|
|   ;
|   return 0;
| }
configure:2961: error: in
`/OE/shr-core/tmp/work/x86_64-linux/elfutils-native-0.148-r1/elfutils-0.148':
configure:2965: error: C compiler cannot create executables
See `config.log' for more details.

Regards,

>
>  do_configure_prepend() {
> --
> 1.7.4.1
>
> _______________________________________________
> poky mailing list
> poky@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/poky
>

Re: [PATCH 1/7] elfutils: fix builds with gcc 4.6

[Richard Purdie]

On Mon, 2011-04-04 at 23:37 +0200, Martin Jansa wrote:
> On Fri, Apr 1, 2011 at 3:40 PM, Joshua Lock <josh@linux.intel.com> wrote:
> > From: Joshua Lock <josh@linux.intel.com>
> >
> > gcc 4.6 (as used in Fedora 15) adds some extra warnings which are
> > included with Werror. The new unused-but-set variable warning causes
> > an error in libasm of elfutils. Work around this by removing
> > unused-but-set from Werror.
> >
> > Signed-off-by: Joshua Lock <josh@linux.intel.com>
> > ---
> >  meta/recipes-devtools/elfutils/elfutils_0.148.bb |    4 ++++
> >  1 files changed, 4 insertions(+), 0 deletions(-)
> >
> > diff --git a/meta/recipes-devtools/elfutils/elfutils_0.148.bb b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> > index b2f700e..c395be8 100644
> > --- a/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> > +++ b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> > @@ -40,6 +40,10 @@ SRC_URI += "\
> >
> >  inherit autotools
> >
> > +# GCC 4.6.0 raises an unused-but-set warning in libasm, for now remove
> > +# this warning from Werror
> > +CFLAGS_virtclass-native += "-Wno-error=unused-but-set-variable"
> > +
> >  EXTRA_OECONF = "--program-prefix=eu-"
> 
> this breaks all other distros not yet using gcc-4.6..
> 
> 
> configure:2892: checking whether the C compiler works
> configure:2914: gcc -Wno-error=unused-but-set-variable
> -isystem/OE/shr-core/tmp/sysroots/x86_64-linux/usr/include
> -L/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
> -Wl,-rpath-link,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
> -Wl,-rpath,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib -Wl,-O1
> conftest.c  >&5
> cc1: error: -Werror=unused-but-set-variable: No option -Wunused-but-set-variable
> configure:2918: $? = 1
> configure:2956: result: no
> configure: failed program was:
> | /* confdefs.h */
> | #define PACKAGE_NAME "Red Hat elfutils"
> | #define PACKAGE_TARNAME "elfutils"
> | #define PACKAGE_VERSION "0.148"
> | #define PACKAGE_STRING "Red Hat elfutils 0.148"
> | #define PACKAGE_BUGREPORT "http://bugzilla.redhat.com/bugzilla/"
> | #define PACKAGE_URL ""
> | #define PACKAGE "elfutils"
> | #define VERSION "0.148"
> | #define MODVERSION "Build on jama 2011-04-04T23:34:25+0200"
> | /* end confdefs.h.  */
> |
> | int
> | main ()
> | {
> |
> |   ;
> |   return 0;
> | }
> configure:2961: error: in
> `/OE/shr-core/tmp/work/x86_64-linux/elfutils-native-0.148-r1/elfutils-0.148':
> configure:2965: error: C compiler cannot create executables
> See `config.log' for more details.

Ouch, I'm going to revert this, we'll need an alternative fix. Thanks
for reporting!

Cheers,

Richard

Re: [PATCH 4/7] xorg-[lib-common|proto-common]: disable use of fop document generation

[Richard Purdie]

On Mon, 2011-04-04 at 23:20 +0200, Martin Jansa wrote:
> On Fri, Apr 1, 2011 at 3:38 PM, Joshua Lock <josh@linux.intel.com> wrote:
> > From: Joshua Lock <josh@linux.intel.com>
> | checking for xmlto... no
> | configure: WARNING: xmlto not found - documentation targets will be skipped
> | configure: error: --with-fop expects 'yes' or 'no'
> | FATAL: oe_runconf failed
> | ERROR: Function 'do_configure' failed (see
> /OE/shr-core/tmp/work/x86_64-linux/xproto-native-1_7.0.20-r0/temp/log.do_configure.29076
> for further information)
> NOTE: package xproto-native-1_7.0.20-r0: task do_configure: Failed
> ERROR: Task 957
> (virtual:native:/OE/shr-core/openembedded-core/meta/recipes-graphics/xorg-proto/xproto_7.0.20.bb,
> do_configure) failed with exit code '1'
> 
> because there is
> --with-fop=no--enable-specs=no
> 
> so you had to put space after no, or before this
> 
> OE nokia900@shr ~/shr-core $ grep OECONF
> openembedded-core/meta/recipes-graphics/xorg-proto/xproto_7.0.20.bb
> EXTRA_OECONF_append = "--enable-specs=no"

Thanks, I'll fix this.

Cheers,

Richard

Re: [PATCH 1/7] elfutils: fix builds with gcc 4.6

[Joshua Lock]

On Mon, 2011-04-04 at 23:46 +0100, Richard Purdie wrote:
> On Mon, 2011-04-04 at 23:37 +0200, Martin Jansa wrote:
> > On Fri, Apr 1, 2011 at 3:40 PM, Joshua Lock <josh@linux.intel.com> wrote:
> > > From: Joshua Lock <josh@linux.intel.com>
> > >
> > > gcc 4.6 (as used in Fedora 15) adds some extra warnings which are
> > > included with Werror. The new unused-but-set variable warning causes
> > > an error in libasm of elfutils. Work around this by removing
> > > unused-but-set from Werror.
> > >
> > > Signed-off-by: Joshua Lock <josh@linux.intel.com>
> > > ---
> > >  meta/recipes-devtools/elfutils/elfutils_0.148.bb |    4 ++++
> > >  1 files changed, 4 insertions(+), 0 deletions(-)
> > >
> > > diff --git a/meta/recipes-devtools/elfutils/elfutils_0.148.bb b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> > > index b2f700e..c395be8 100644
> > > --- a/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> > > +++ b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> > > @@ -40,6 +40,10 @@ SRC_URI += "\
> > >
> > >  inherit autotools
> > >
> > > +# GCC 4.6.0 raises an unused-but-set warning in libasm, for now remove
> > > +# this warning from Werror
> > > +CFLAGS_virtclass-native += "-Wno-error=unused-but-set-variable"
> > > +
> > >  EXTRA_OECONF = "--program-prefix=eu-"
> > 
> > this breaks all other distros not yet using gcc-4.6..
> > 
> > 
> > configure:2892: checking whether the C compiler works
> > configure:2914: gcc -Wno-error=unused-but-set-variable
> > -isystem/OE/shr-core/tmp/sysroots/x86_64-linux/usr/include
> > -L/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
> > -Wl,-rpath-link,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
> > -Wl,-rpath,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib -Wl,-O1
> > conftest.c  >&5
> > cc1: error: -Werror=unused-but-set-variable: No option -Wunused-but-set-variable
> > configure:2918: $? = 1
> > configure:2956: result: no
> > configure: failed program was:
> > | /* confdefs.h */
> > | #define PACKAGE_NAME "Red Hat elfutils"
> > | #define PACKAGE_TARNAME "elfutils"
> > | #define PACKAGE_VERSION "0.148"
> > | #define PACKAGE_STRING "Red Hat elfutils 0.148"
> > | #define PACKAGE_BUGREPORT "http://bugzilla.redhat.com/bugzilla/"
> > | #define PACKAGE_URL ""
> > | #define PACKAGE "elfutils"
> > | #define VERSION "0.148"
> > | #define MODVERSION "Build on jama 2011-04-04T23:34:25+0200"
> > | /* end confdefs.h.  */
> > |
> > | int
> > | main ()
> > | {
> > |
> > |   ;
> > |   return 0;
> > | }
> > configure:2961: error: in
> > `/OE/shr-core/tmp/work/x86_64-linux/elfutils-native-0.148-r1/elfutils-0.148':
> > configure:2965: error: C compiler cannot create executables
> > See `config.log' for more details.

Sorry about this Martin. Thanks for catching and reporting.

> 
> Ouch, I'm going to revert this, we'll need an alternative fix. Thanks
> for reporting!

I've worked up an alternative fix and sent a pull request to oe-core. I
took the approach of consolidating several patches from the upstream
source repository which remove the unused variables.

I consolidated them into a single patch so that they are easier to drop
once we update elfutils to 0.152 or higher.

Regards,
Joshua
-- 
Joshua Lock
        Yocto Build System Monkey
        Intel Open Source Technology Centre

Re: [PATCH 1/7] elfutils: fix builds with gcc 4.6

[Khem Raj]

On Wed, Apr 6, 2011 at 4:16 AM, Joshua Lock <josh@linux.intel.com> wrote:
> On Mon, 2011-04-04 at 23:46 +0100, Richard Purdie wrote:
>> On Mon, 2011-04-04 at 23:37 +0200, Martin Jansa wrote:
>> > On Fri, Apr 1, 2011 at 3:40 PM, Joshua Lock <josh@linux.intel.com> wrote:
>> > > From: Joshua Lock <josh@linux.intel.com>
>> > >
>> > > gcc 4.6 (as used in Fedora 15) adds some extra warnings which are
>> > > included with Werror. The new unused-but-set variable warning causes
>> > > an error in libasm of elfutils. Work around this by removing
>> > > unused-but-set from Werror.
>> > >
>> > > Signed-off-by: Joshua Lock <josh@linux.intel.com>
>> > > ---
>> > >  meta/recipes-devtools/elfutils/elfutils_0.148.bb |    4 ++++
>> > >  1 files changed, 4 insertions(+), 0 deletions(-)
>> > >
>> > > diff --git a/meta/recipes-devtools/elfutils/elfutils_0.148.bb b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
>> > > index b2f700e..c395be8 100644
>> > > --- a/meta/recipes-devtools/elfutils/elfutils_0.148.bb
>> > > +++ b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
>> > > @@ -40,6 +40,10 @@ SRC_URI += "\
>> > >
>> > >  inherit autotools
>> > >
>> > > +# GCC 4.6.0 raises an unused-but-set warning in libasm, for now remove
>> > > +# this warning from Werror
>> > > +CFLAGS_virtclass-native += "-Wno-error=unused-but-set-variable"
>> > > +
>> > >  EXTRA_OECONF = "--program-prefix=eu-"
>> >
>> > this breaks all other distros not yet using gcc-4.6..
>> >
>> >
>> > configure:2892: checking whether the C compiler works
>> > configure:2914: gcc -Wno-error=unused-but-set-variable
>> > -isystem/OE/shr-core/tmp/sysroots/x86_64-linux/usr/include
>> > -L/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
>> > -Wl,-rpath-link,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
>> > -Wl,-rpath,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib -Wl,-O1
>> > conftest.c  >&5
>> > cc1: error: -Werror=unused-but-set-variable: No option -Wunused-but-set-variable
>> > configure:2918: $? = 1
>> > configure:2956: result: no
>> > configure: failed program was:
>> > | /* confdefs.h */
>> > | #define PACKAGE_NAME "Red Hat elfutils"
>> > | #define PACKAGE_TARNAME "elfutils"
>> > | #define PACKAGE_VERSION "0.148"
>> > | #define PACKAGE_STRING "Red Hat elfutils 0.148"
>> > | #define PACKAGE_BUGREPORT "http://bugzilla.redhat.com/bugzilla/"
>> > | #define PACKAGE_URL ""
>> > | #define PACKAGE "elfutils"
>> > | #define VERSION "0.148"
>> > | #define MODVERSION "Build on jama 2011-04-04T23:34:25+0200"
>> > | /* end confdefs.h.  */
>> > |
>> > | int
>> > | main ()
>> > | {
>> > |
>> > |   ;
>> > |   return 0;
>> > | }
>> > configure:2961: error: in
>> > `/OE/shr-core/tmp/work/x86_64-linux/elfutils-native-0.148-r1/elfutils-0.148':
>> > configure:2965: error: C compiler cannot create executables
>> > See `config.log' for more details.
>
> Sorry about this Martin. Thanks for catching and reporting.
>
>>
>> Ouch, I'm going to revert this, we'll need an alternative fix. Thanks
>> for reporting!
>
> I've worked up an alternative fix and sent a pull request to oe-core. I
> took the approach of consolidating several patches from the upstream
> source repository which remove the unused variables.
>
> I consolidated them into a single patch so that they are easier to drop
> once we update elfutils to 0.152 or higher.
>

Why not update to 0.152 then

> Regards,
> Joshua
> --
> Joshua Lock
>        Yocto Build System Monkey
>        Intel Open Source Technology Centre
>
> _______________________________________________
> poky mailing list
> poky@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/poky
>

Re: [PATCH 1/7] elfutils: fix builds with gcc 4.6

[Joshua Lock]

On Wed, 2011-04-06 at 10:06 -0700, Khem Raj wrote:
> On Wed, Apr 6, 2011 at 4:16 AM, Joshua Lock <josh@linux.intel.com> wrote:
> > On Mon, 2011-04-04 at 23:46 +0100, Richard Purdie wrote:
> >> On Mon, 2011-04-04 at 23:37 +0200, Martin Jansa wrote:
> >> > On Fri, Apr 1, 2011 at 3:40 PM, Joshua Lock <josh@linux.intel.com> wrote:
> >> > > From: Joshua Lock <josh@linux.intel.com>
> >> > >
> >> > > gcc 4.6 (as used in Fedora 15) adds some extra warnings which are
> >> > > included with Werror. The new unused-but-set variable warning causes
> >> > > an error in libasm of elfutils. Work around this by removing
> >> > > unused-but-set from Werror.
> >> > >
> >> > > Signed-off-by: Joshua Lock <josh@linux.intel.com>
> >> > > ---
> >> > >  meta/recipes-devtools/elfutils/elfutils_0.148.bb |    4 ++++
> >> > >  1 files changed, 4 insertions(+), 0 deletions(-)
> >> > >
> >> > > diff --git a/meta/recipes-devtools/elfutils/elfutils_0.148.bb b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> >> > > index b2f700e..c395be8 100644
> >> > > --- a/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> >> > > +++ b/meta/recipes-devtools/elfutils/elfutils_0.148.bb
> >> > > @@ -40,6 +40,10 @@ SRC_URI += "\
> >> > >
> >> > >  inherit autotools
> >> > >
> >> > > +# GCC 4.6.0 raises an unused-but-set warning in libasm, for now remove
> >> > > +# this warning from Werror
> >> > > +CFLAGS_virtclass-native += "-Wno-error=unused-but-set-variable"
> >> > > +
> >> > >  EXTRA_OECONF = "--program-prefix=eu-"
> >> >
> >> > this breaks all other distros not yet using gcc-4.6..
> >> >
> >> >
> >> > configure:2892: checking whether the C compiler works
> >> > configure:2914: gcc -Wno-error=unused-but-set-variable
> >> > -isystem/OE/shr-core/tmp/sysroots/x86_64-linux/usr/include
> >> > -L/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
> >> > -Wl,-rpath-link,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib
> >> > -Wl,-rpath,/OE/shr-core/tmp/sysroots/x86_64-linux/usr/lib -Wl,-O1
> >> > conftest.c  >&5
> >> > cc1: error: -Werror=unused-but-set-variable: No option -Wunused-but-set-variable
> >> > configure:2918: $? = 1
> >> > configure:2956: result: no
> >> > configure: failed program was:
> >> > | /* confdefs.h */
> >> > | #define PACKAGE_NAME "Red Hat elfutils"
> >> > | #define PACKAGE_TARNAME "elfutils"
> >> > | #define PACKAGE_VERSION "0.148"
> >> > | #define PACKAGE_STRING "Red Hat elfutils 0.148"
> >> > | #define PACKAGE_BUGREPORT "http://bugzilla.redhat.com/bugzilla/"
> >> > | #define PACKAGE_URL ""
> >> > | #define PACKAGE "elfutils"
> >> > | #define VERSION "0.148"
> >> > | #define MODVERSION "Build on jama 2011-04-04T23:34:25+0200"
> >> > | /* end confdefs.h.  */
> >> > |
> >> > | int
> >> > | main ()
> >> > | {
> >> > |
> >> > |   ;
> >> > |   return 0;
> >> > | }
> >> > configure:2961: error: in
> >> > `/OE/shr-core/tmp/work/x86_64-linux/elfutils-native-0.148-r1/elfutils-0.148':
> >> > configure:2965: error: C compiler cannot create executables
> >> > See `config.log' for more details.
> >
> > Sorry about this Martin. Thanks for catching and reporting.
> >
> >>
> >> Ouch, I'm going to revert this, we'll need an alternative fix. Thanks
> >> for reporting!
> >
> > I've worked up an alternative fix and sent a pull request to oe-core. I
> > took the approach of consolidating several patches from the upstream
> > source repository which remove the unused variables.
> >
> > I consolidated them into a single patch so that they are easier to drop
> > once we update elfutils to 0.152 or higher.
> >
> 
> Why not update to 0.152 then

The patch series is intended for a Bernard point release, the Bernard
release is already out with eflutils 0.148.

Regards,
Joshua
-- 
Joshua Lock
        Yocto Build System Monkey
        Intel Open Source Technology Centre

Re: [PATCH 1/7] elfutils: fix builds with gcc 4.6

[Richard Purdie]

On Fri, 2011-04-08 at 13:28 +0100, Joshua Lock wrote:
> On Wed, 2011-04-06 at 10:06 -0700, Khem Raj wrote:
> > On Wed, Apr 6, 2011 at 4:16 AM, Joshua Lock <josh@linux.intel.com> wrote:
> > > I've worked up an alternative fix and sent a pull request to oe-core. I
> > > took the approach of consolidating several patches from the upstream
> > > source repository which remove the unused variables.
> > >
> > > I consolidated them into a single patch so that they are easier to drop
> > > once we update elfutils to 0.152 or higher.
> > >
> > 
> > Why not update to 0.152 then
> 
> The patch series is intended for a Bernard point release, the Bernard
> release is already out with eflutils 0.148.

I think this is a good approach given the issues we know we're going to
have with Bernard. I'll obviously take a version upgrade patch too :)

Cheers,

Richard