[PATCH 0/3] misc fixups

[PATCH 0/3] misc fixups

[Bernhard Reutner-Fischer]

Hi,

The last patch (gcc-cross) asks for comments if and how the package
version number should be improved or look like.
Please apply the rest as-is (omit that PV-touching hunk if in doubt).
thanks,

Bernhard Reutner-Fischer (3):
  minimal: commentary typo fix
  micro: get PREFERRED_BINUTILS from sane-toolchain
  gcc-cross: fix PV and make it compile

 conf/distro/micro.conf                             |    1 -
 conf/distro/minimal-uclibc.conf                    |    2 +-
 recipes/gcc/gcc-svn.inc                            |    8 +-
 recipes/gcc/gcc-svn/gcc-flags-for-build.patch      |   26 --
 .../gcc-uclibc-locale-ctype_touplow_t.patch        |   99 +++++---
 .../powerpc32-c++-pch-long-double-128.patch        |  255 ++++++++++++++++++++
 6 files changed, 323 insertions(+), 68 deletions(-)
 create mode 100644 recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch

[PATCH 1/3] minimal: commentary typo fix

[Bernhard Reutner-Fischer]

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
---
 conf/distro/minimal-uclibc.conf |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/conf/distro/minimal-uclibc.conf b/conf/distro/minimal-uclibc.conf
index 44805cf..3347811 100644
--- a/conf/distro/minimal-uclibc.conf
+++ b/conf/distro/minimal-uclibc.conf
@@ -42,5 +42,5 @@ DISTRO_FEATURES += "eabi"
 DISTRO_FEATURES += "pam"
 # large file support is needed for libsndfile
 DISTRO_FEATURES += "largefile"
-# glib-2.0 nneds ipv6 and there is no knob to control it
+# glib-2.0 needs ipv6 and there is no knob to control it
 DISTRO_FEATURES += "ipv4 ipv6"
-- 
1.7.1

[PATCH 2/3] micro: get PREFERRED_BINUTILS from sane-toolchain

[Bernhard Reutner-Fischer]

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
---
 conf/distro/micro.conf |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/conf/distro/micro.conf b/conf/distro/micro.conf
index 9885425..3334692 100644
--- a/conf/distro/micro.conf
+++ b/conf/distro/micro.conf
@@ -61,7 +61,6 @@ MACHINE_KERNEL_VERSION = "2.6"
 # TOOLCHAIN
 #############################################################################
 LIBC ?= "eglibc"
-PREFERRED_BINUTILS = "2.19.51"
 PREFERRED_VERSION_glibc             ?= "2.10.1"
 PREFERRED_VERSION_glibc-initial     ?= "2.10.1"
 PREFERRED_ARM_INSTRUCTION_SET ?= "thumb"
-- 
1.7.1

[PATCH 3/3] gcc-cross: fix PV and make it compile

[Bernhard Reutner-Fischer]

- using PREFERRED_GCC_VERSION = "svn" didn't work since
  gcc-svn.inc messed with PV
  TODO: tweak package version string?
        now "...svn-r04.6+svnr162209" could be made neater, perhaps
- disable libstdcxx-pch for ppc
  pch1a_output would need to be compiled with -mlong-double-128
  to avoid
    libstdc++-v3/include/precompiled/stdc++.h -o \
	powerpc-oe-linux-uclibc/bits/stdc++.h.gch/O2g.gch
    [snip]
    libstdc++-v3/include/tr1_impl/cmath:153:11: error: '::acoshl' has not been declared
  install proper guards and furthermore disable libstdcxx-pch for now

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
---
 recipes/gcc/gcc-svn.inc                            |    8 +-
 recipes/gcc/gcc-svn/gcc-flags-for-build.patch      |   26 --
 .../gcc-uclibc-locale-ctype_touplow_t.patch        |   99 +++++---
 .../powerpc32-c++-pch-long-double-128.patch        |  255 ++++++++++++++++++++
 4 files changed, 322 insertions(+), 66 deletions(-)
 create mode 100644 recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch

diff --git a/recipes/gcc/gcc-svn.inc b/recipes/gcc/gcc-svn.inc
index 2e24d63..28307d2 100644
--- a/recipes/gcc/gcc-svn.inc
+++ b/recipes/gcc/gcc-svn.inc
@@ -1,7 +1,7 @@
 DEFAULT_PREFERENCE = "-999"
 ARM_INSTRUCTION_SET = "arm"
 
-SRCREV = "158653"
+SRCREV = "162223"
 
 require gcc-common.inc
 
@@ -9,7 +9,7 @@ DEPENDS = "mpfr gmp libmpc"
 NATIVEDEPS = "mpfr-native gmp-native libmpc-native"
 
 BINV = "4.6.0"
-PV = "4.6+svnr${SRCPV}"
+PR_append = "4.6+svnr${SRCPV}"
 
 FILESPATHPKG .= ":gcc-svn"
 
@@ -19,6 +19,7 @@ SRC_URI = "svn://gcc.gnu.org/svn/gcc;module=trunk \
            file://gcc-uclibc-locale-ctype_touplow_t.patch \
 	   file://cache-amnesia.patch \
 	   file://gcc-flags-for-build.patch \
+	   file://powerpc32-c++-pch-long-double-128.patch \
 	  "
 
 # Language Overrides
@@ -42,3 +43,6 @@ do_unpack_append() {
 EXTRA_OECONF_append_linux-uclibc = " --disable-decimal-float "
 EXTRA_OECONF_append_linux-uclibceabi = " --disable-decimal-float "
 EXTRA_OECONF_append_linux-uclibcspe = " --disable-decimal-float "
+# powerpc32 would need to build libstdc++-v3/include/cmath with -mlong-double-128
+# just disable PCH for ppc
+EXTRA_OECONF_append_powerpc-linux-uclibc = " --disable-libstdcxx-pch "
diff --git a/recipes/gcc/gcc-svn/gcc-flags-for-build.patch b/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
index 98b3ecf..0a0da9a 100644
--- a/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
+++ b/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
@@ -98,24 +98,6 @@ Index: gcc-4.5+svnr155680/gcc/configure
  fi
  
  # Expand extra_headers to include complete path.
-@@ -17037,7 +17040,7 @@ else
-   lt_dlunknown=0; lt_dlno_uscore=1; lt_dlneed_uscore=2
-   lt_status=$lt_dlunknown
-   cat > conftest.$ac_ext <<_LT_EOF
--#line 17040 "configure"
-+#line 17043 "configure"
- #include "confdefs.h"
- 
- #if HAVE_DLFCN_H
-@@ -17143,7 +17146,7 @@ else
-   lt_dlunknown=0; lt_dlno_uscore=1; lt_dlneed_uscore=2
-   lt_status=$lt_dlunknown
-   cat > conftest.$ac_ext <<_LT_EOF
--#line 17146 "configure"
-+#line 17149 "configure"
- #include "confdefs.h"
- 
- #if HAVE_DLFCN_H
 Index: gcc-4.5+svnr155680/Makefile.tpl
 ===================================================================
 --- gcc-4.5+svnr155680.orig/Makefile.tpl	2010-01-13 01:02:22.603540396 -0800
@@ -148,14 +130,6 @@ Index: gcc-4.5+svnr155680/configure
    CXXFLAGS_FOR_BUILD=${CXXFLAGS_FOR_BUILD-${CXXFLAGS}}
    LDFLAGS_FOR_BUILD=${LDFLAGS_FOR_BUILD-${LDFLAGS}}
  fi
-@@ -7940,6 +7942,7 @@ done
- 
- 
- 
-+
- # Generate default definitions for YACC, M4, LEX and other programs that run
- # on the build machine.  These are used if the Makefile can't locate these
- # programs in objdir.
 Index: gcc-4.5+svnr155680/configure.ac
 ===================================================================
 --- gcc-4.5+svnr155680.orig/configure.ac	2010-01-13 01:02:22.651540130 -0800
diff --git a/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch b/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
index ffb37d2..9cae744 100644
--- a/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
+++ b/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
@@ -1,52 +1,75 @@
-Index: gcc-4.4.1/libstdc++-v3/config/locale/generic/c_locale.h
-===================================================================
---- gcc-4.4.1.orig/libstdc++-v3/config/locale/generic/c_locale.h	2009-08-06 23:38:32.398265633 -0700
-+++ gcc-4.4.1/libstdc++-v3/config/locale/generic/c_locale.h	2009-08-06 23:41:09.778242281 -0700
-@@ -41,12 +41,17 @@
- 
- #include <clocale>
- #include <cstddef>
-+#include <features.h> 
-+#include <ctype.h> 
- 
+diff --git a/libstdc++-v3/config/locale/generic/c_locale.h b/libstdc++-v3/config/locale/generic/c_locale.h
+index 7d685b8..4f69857 100644
+--- a/libstdc++-v3/config/locale/generic/c_locale.h
++++ b/libstdc++-v3/config/locale/generic/c_locale.h
+@@ -44,8 +44,11 @@
  #define _GLIBCXX_NUM_CATEGORIES 0
  
  _GLIBCXX_BEGIN_NAMESPACE(std)
 -
--  typedef int*			__c_locale;
-+#ifdef __UCLIBC__ 
-+   typedef __ctype_touplow_t*   __c_locale; 
-+#else 
-+   typedef int*         __c_locale; 
-+#endif 
++#ifdef __UCLIBC__
++  typedef __ctype_touplow_t* __c_locale;
++#else
+   typedef int*			__c_locale;
++#endif
  
    // Convert numeric value of type double and long double to string and
    // return length of string.  If vsnprintf is available use it, otherwise
-Index: gcc-4.4.1/libstdc++-v3/config/os/gnu-linux/ctype_base.h
-===================================================================
---- gcc-4.4.1.orig/libstdc++-v3/config/os/gnu-linux/ctype_base.h	2009-08-06 23:42:27.402242608 -0700
-+++ gcc-4.4.1/libstdc++-v3/config/os/gnu-linux/ctype_base.h	2009-08-06 23:43:44.842241959 -0700
-@@ -33,14 +33,21 @@
-  */
- 
- // Information as gleaned from /usr/include/ctype.h
--  
-+
-+#include <features.h> 
-+#include <ctype.h> 
-+
- _GLIBCXX_BEGIN_NAMESPACE(std)
- 
-   /// @brief  Base class for ctype.
+diff --git a/libstdc++-v3/config/os/gnu-linux/ctype_base.h b/libstdc++-v3/config/os/gnu-linux/ctype_base.h
+index ff1f157..e88f569 100644
+--- a/libstdc++-v3/config/os/gnu-linux/ctype_base.h
++++ b/libstdc++-v3/config/os/gnu-linux/ctype_base.h
+@@ -40,7 +40,11 @@ _GLIBCXX_BEGIN_NAMESPACE(std)
    struct ctype_base
    {
      // Non-standard typedefs.
--    typedef const int* 		__to_type;
-+#ifdef __UCLIBC__ 
-+    typedef const __ctype_touplow_t*   __to_type; 
-+#else 
-+    typedef const int*         __to_type; 
++#ifdef __UCLIBC__
++    typedef const __ctype_touplow_t* __to_type;
++#else
+     typedef const int* 		__to_type;
 +#endif
  
      // NB: Offsets into ctype<char>::_M_table force a particular size
      // on the mask type. Because of this, we don't use an enum.
+diff --git a/libstdc++-v3/include/c/clocale b/libstdc++-v3/include/c/clocale
+index 7ea60f0..619f128 100644
+--- a/libstdc++-v3/include/c/clocale
++++ b/libstdc++-v3/include/c/clocale
+@@ -32,5 +32,8 @@
+ #pragma GCC system_header
+ 
+ #include_next <locale.h>
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
+ 
+ #endif
+diff --git a/libstdc++-v3/include/c_global/clocale b/libstdc++-v3/include/c_global/clocale
+index 7cf42e2..cd9b37a 100644
+--- a/libstdc++-v3/include/c_global/clocale
++++ b/libstdc++-v3/include/c_global/clocale
+@@ -42,6 +42,9 @@
+ 
+ #include <bits/c++config.h>
+ #include <locale.h>
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
+ 
+ #ifndef _GLIBCXX_CLOCALE
+ #define _GLIBCXX_CLOCALE 1
+diff --git a/libstdc++-v3/include/c_std/clocale b/libstdc++-v3/include/c_std/clocale
+index 36d8bd7..7905d5d 100644
+--- a/libstdc++-v3/include/c_std/clocale
++++ b/libstdc++-v3/include/c_std/clocale
+@@ -45,6 +45,10 @@
+ #include <bits/c++config.h>
+ #include <locale.h>
+ 
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
++
+ // Get rid of those macros defined in <locale.h> in lieu of real functions.
+ #undef setlocale
+ #undef localeconv
diff --git a/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch b/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch
new file mode 100644
index 0000000..d20e3d0
--- /dev/null
+++ b/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch
@@ -0,0 +1,255 @@
+on powerpc32 cmath has to be compiled with -mlong-double-128
+
+diff --git a/libstdc++-v3/include/tr1_impl/cmath b/libstdc++-v3/include/tr1_impl/cmath
+index 5807380..f395a0d 100644
+--- a/libstdc++-v3/include/tr1_impl/cmath
++++ b/libstdc++-v3/include/tr1_impl/cmath
+@@ -150,143 +150,213 @@ _GLIBCXX_BEGIN_NAMESPACE_TR1
+   // functions
+   using ::acosh;
+   using ::acoshf;
+-  using ::acoshl;
++#ifdef __LONG_DOUBLE_128__
++   using ::acoshl;
++#endif
+ 
+   using ::asinh;
+   using ::asinhf;
+-  using ::asinhl;
++#ifdef __LONG_DOUBLE_128__
++   using ::asinhl;
++#endif
+ 
+   using ::atanh;
+   using ::atanhf;
+-  using ::atanhl;
++#ifdef __LONG_DOUBLE_128__
++   using ::atanhl;
++#endif
+ 
+   using ::cbrt;
+   using ::cbrtf;
+-  using ::cbrtl;
++#ifdef __LONG_DOUBLE_128__
++   using ::cbrtl;
++#endif
+ 
+   using ::copysign;
+   using ::copysignf;
+-  using ::copysignl;
++#ifdef __LONG_DOUBLE_128__
++   using ::copysignl;
++#endif
+ 
+   using ::erf;
+   using ::erff;
+-  using ::erfl;
++#ifdef __LONG_DOUBLE_128__
++   using ::erfl;
++#endif
+ 
+   using ::erfc;
+   using ::erfcf;
+-  using ::erfcl;
++#ifdef __LONG_DOUBLE_128__
++   using ::erfcl;
++#endif
+ 
+   using ::exp2;
+   using ::exp2f;
+-  using ::exp2l;
++#ifdef __LONG_DOUBLE_128__
++   using ::exp2l;
++#endif
+ 
+   using ::expm1;
+   using ::expm1f;
+-  using ::expm1l;
++#ifdef __LONG_DOUBLE_128__
++   using ::expm1l;
++#endif
+ 
+   using ::fdim;
+   using ::fdimf;
+-  using ::fdiml;
++#ifdef __LONG_DOUBLE_128__
++   using ::fdiml;
++#endif
+ 
+   using ::fma;
+   using ::fmaf;
+-  using ::fmal;
++#ifdef __LONG_DOUBLE_128__
++   using ::fmal;
++#endif
+ 
+   using ::fmax;
+   using ::fmaxf;
+-  using ::fmaxl;
++#ifdef __LONG_DOUBLE_128__
++   using ::fmaxl;
++#endif
+ 
+   using ::fmin;
+   using ::fminf;
+-  using ::fminl;
++#ifdef __LONG_DOUBLE_128__
++   using ::fminl;
++#endif
+ 
+   using ::hypot;
+   using ::hypotf;
+-  using ::hypotl;
++#ifdef __LONG_DOUBLE_128__
++   using ::hypotl;
++#endif
+ 
+   using ::ilogb;
+   using ::ilogbf;
+-  using ::ilogbl;
++#ifdef __LONG_DOUBLE_128__
++   using ::ilogbl;
++#endif
+ 
+   using ::lgamma;
+   using ::lgammaf;
+-  using ::lgammal;
++#ifdef __LONG_DOUBLE_128__
++   using ::lgammal;
++#endif
+ 
+   using ::llrint;
+   using ::llrintf;
+-  using ::llrintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::llrintl;
++#endif
+ 
+   using ::llround;
+   using ::llroundf;
+-  using ::llroundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::llroundl;
++#endif
+ 
+   using ::log1p;
+   using ::log1pf;
+-  using ::log1pl;
++#ifdef __LONG_DOUBLE_128__
++   using ::log1pl;
++#endif
+ 
+   using ::log2;
+   using ::log2f;
+-  using ::log2l;
++#ifdef __LONG_DOUBLE_128__
++   using ::log2l;
++#endif
+ 
+   using ::logb;
+   using ::logbf;
+-  using ::logbl;
++#ifdef __LONG_DOUBLE_128__
++   using ::logbl;
++#endif
+ 
+   using ::lrint;
+   using ::lrintf;
+-  using ::lrintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::lrintl;
++#endif
+ 
+   using ::lround;
+   using ::lroundf;
+-  using ::lroundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::lroundl;
++#endif
+ 
+   using ::nan;
+   using ::nanf;
+-  using ::nanl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nanl;
++#endif
+ 
+   using ::nearbyint;
+   using ::nearbyintf;
+-  using ::nearbyintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nearbyintl;
++#endif
+ 
+   using ::nextafter;
+   using ::nextafterf;
+-  using ::nextafterl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nextafterl;
++#endif
+ 
+   using ::nexttoward;
+   using ::nexttowardf;
+-  using ::nexttowardl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nexttowardl;
++#endif
+ 
+   using ::remainder;
+   using ::remainderf;
+-  using ::remainderl;
++#ifdef __LONG_DOUBLE_128__
++   using ::remainderl;
++#endif
+ 
+   using ::remquo;
+   using ::remquof;
+-  using ::remquol;
++#ifdef __LONG_DOUBLE_128__
++   using ::remquol;
++#endif
+ 
+   using ::rint;
+   using ::rintf;
+-  using ::rintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::rintl;
++#endif
+ 
+   using ::round;
+   using ::roundf;
+-  using ::roundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::roundl;
++#endif
+ 
+   using ::scalbln;
+   using ::scalblnf;
+-  using ::scalblnl;
++#ifdef __LONG_DOUBLE_128__
++   using ::scalblnl;
++#endif
+ 
+   using ::scalbn;
+   using ::scalbnf;
+-  using ::scalbnl;
++#ifdef __LONG_DOUBLE_128__
++   using ::scalbnl;
++#endif
+ 
+   using ::tgamma;
+   using ::tgammaf;
+-  using ::tgammal;
++#ifdef __LONG_DOUBLE_128__
++   using ::tgammal;
++#endif
+ 
+   using ::trunc;
+   using ::truncf;
+-  using ::truncl;
++#ifdef __LONG_DOUBLE_128__
++   using ::truncl;
++#endif
+ 
+ #endif
+ 
-- 
1.7.1

Re: [PATCH 3/3] gcc-cross: fix PV and make it compile

[Phil Blundell]

On Thu, 2010-07-15 at 17:25 +0200, Bernhard Reutner-Fischer wrote:
> -PV = "4.6+svnr${SRCPV}"
> +PR_append = "4.6+svnr${SRCPV}"

Munging PR like this is fairly ugly.  You might consider bashing PKGV
instead.

p.

Re: [PATCH 3/3] gcc-cross: fix PV and make it compile

[Martin Jansa]

On Thu, Jul 15, 2010 at 04:48:04PM +0100, Phil Blundell wrote:
> On Thu, 2010-07-15 at 17:25 +0200, Bernhard Reutner-Fischer wrote:
> > -PV = "4.6+svnr${SRCPV}"
> > +PR_append = "4.6+svnr${SRCPV}"
> 
> Munging PR like this is fairly ugly.  You might consider bashing PKGV
> instead.

PREFERRED_GCC_VERSION = "4.6+svnr%' should also work with bitbake-1.10 and newer

Regards,

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com

[PATCH][v2] gcc-cross: fix PV and make it compile

[Bernhard Reutner-Fischer]

- using PREFERRED_GCC_VERSION = "svn" didn't work since
  gcc-svn.inc messed with PV
- disable libstdcxx-pch for ppc
  pch1a_output would need to be compiled with -mlong-double-128
  to avoid
    libstdc++-v3/include/precompiled/stdc++.h -o \
	powerpc-oe-linux-uclibc/bits/stdc++.h.gch/O2g.gch
    [snip]
    libstdc++-v3/include/tr1_impl/cmath:153:11: error: '::acoshl' has not been declared
  install proper guards and furthermore disable libstdcxx-pch for now

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
---
 recipes/gcc/gcc-svn.inc                            |   10 +-
 recipes/gcc/gcc-svn/gcc-flags-for-build.patch      |   26 --
 .../gcc-uclibc-locale-ctype_touplow_t.patch        |   99 +++++---
 .../powerpc32-c++-pch-long-double-128.patch        |  255 ++++++++++++++++++++
 4 files changed, 323 insertions(+), 67 deletions(-)
 create mode 100644 recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch

diff --git a/recipes/gcc/gcc-svn.inc b/recipes/gcc/gcc-svn.inc
index 2e24d63..40b46d6 100644
--- a/recipes/gcc/gcc-svn.inc
+++ b/recipes/gcc/gcc-svn.inc
@@ -1,7 +1,7 @@
 DEFAULT_PREFERENCE = "-999"
 ARM_INSTRUCTION_SET = "arm"
 
-SRCREV = "158653"
+SRCREV = "162223"
 
 require gcc-common.inc
 
@@ -9,9 +9,9 @@ DEPENDS = "mpfr gmp libmpc"
 NATIVEDEPS = "mpfr-native gmp-native libmpc-native"
 
 BINV = "4.6.0"
-PV = "4.6+svnr${SRCPV}"
+PKGV = "4.6+svnr${SRCPV}"
 
-FILESPATHPKG .= ":gcc-svn"
+FILESPATHPKG .= ":gcc-${PV}"
 
 SRC_URI = "svn://gcc.gnu.org/svn/gcc;module=trunk \
 	   file://gcc-4.3.1-ARCH_FLAGS_FOR_TARGET.patch \
@@ -19,6 +19,7 @@ SRC_URI = "svn://gcc.gnu.org/svn/gcc;module=trunk \
            file://gcc-uclibc-locale-ctype_touplow_t.patch \
 	   file://cache-amnesia.patch \
 	   file://gcc-flags-for-build.patch \
+	   file://powerpc32-c++-pch-long-double-128.patch \
 	  "
 
 # Language Overrides
@@ -42,3 +43,6 @@ do_unpack_append() {
 EXTRA_OECONF_append_linux-uclibc = " --disable-decimal-float "
 EXTRA_OECONF_append_linux-uclibceabi = " --disable-decimal-float "
 EXTRA_OECONF_append_linux-uclibcspe = " --disable-decimal-float "
+# powerpc32 would need to build libstdc++-v3/include/cmath with -mlong-double-128
+# just disable PCH for ppc
+EXTRA_OECONF_append_powerpc-linux-uclibc = " --disable-libstdcxx-pch "
diff --git a/recipes/gcc/gcc-svn/gcc-flags-for-build.patch b/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
index 98b3ecf..0a0da9a 100644
--- a/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
+++ b/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
@@ -98,24 +98,6 @@ Index: gcc-4.5+svnr155680/gcc/configure
  fi
  
  # Expand extra_headers to include complete path.
-@@ -17037,7 +17040,7 @@ else
-   lt_dlunknown=0; lt_dlno_uscore=1; lt_dlneed_uscore=2
-   lt_status=$lt_dlunknown
-   cat > conftest.$ac_ext <<_LT_EOF
--#line 17040 "configure"
-+#line 17043 "configure"
- #include "confdefs.h"
- 
- #if HAVE_DLFCN_H
-@@ -17143,7 +17146,7 @@ else
-   lt_dlunknown=0; lt_dlno_uscore=1; lt_dlneed_uscore=2
-   lt_status=$lt_dlunknown
-   cat > conftest.$ac_ext <<_LT_EOF
--#line 17146 "configure"
-+#line 17149 "configure"
- #include "confdefs.h"
- 
- #if HAVE_DLFCN_H
 Index: gcc-4.5+svnr155680/Makefile.tpl
 ===================================================================
 --- gcc-4.5+svnr155680.orig/Makefile.tpl	2010-01-13 01:02:22.603540396 -0800
@@ -148,14 +130,6 @@ Index: gcc-4.5+svnr155680/configure
    CXXFLAGS_FOR_BUILD=${CXXFLAGS_FOR_BUILD-${CXXFLAGS}}
    LDFLAGS_FOR_BUILD=${LDFLAGS_FOR_BUILD-${LDFLAGS}}
  fi
-@@ -7940,6 +7942,7 @@ done
- 
- 
- 
-+
- # Generate default definitions for YACC, M4, LEX and other programs that run
- # on the build machine.  These are used if the Makefile can't locate these
- # programs in objdir.
 Index: gcc-4.5+svnr155680/configure.ac
 ===================================================================
 --- gcc-4.5+svnr155680.orig/configure.ac	2010-01-13 01:02:22.651540130 -0800
diff --git a/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch b/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
index ffb37d2..9cae744 100644
--- a/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
+++ b/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
@@ -1,52 +1,75 @@
-Index: gcc-4.4.1/libstdc++-v3/config/locale/generic/c_locale.h
-===================================================================
---- gcc-4.4.1.orig/libstdc++-v3/config/locale/generic/c_locale.h	2009-08-06 23:38:32.398265633 -0700
-+++ gcc-4.4.1/libstdc++-v3/config/locale/generic/c_locale.h	2009-08-06 23:41:09.778242281 -0700
-@@ -41,12 +41,17 @@
- 
- #include <clocale>
- #include <cstddef>
-+#include <features.h> 
-+#include <ctype.h> 
- 
+diff --git a/libstdc++-v3/config/locale/generic/c_locale.h b/libstdc++-v3/config/locale/generic/c_locale.h
+index 7d685b8..4f69857 100644
+--- a/libstdc++-v3/config/locale/generic/c_locale.h
++++ b/libstdc++-v3/config/locale/generic/c_locale.h
+@@ -44,8 +44,11 @@
  #define _GLIBCXX_NUM_CATEGORIES 0
  
  _GLIBCXX_BEGIN_NAMESPACE(std)
 -
--  typedef int*			__c_locale;
-+#ifdef __UCLIBC__ 
-+   typedef __ctype_touplow_t*   __c_locale; 
-+#else 
-+   typedef int*         __c_locale; 
-+#endif 
++#ifdef __UCLIBC__
++  typedef __ctype_touplow_t* __c_locale;
++#else
+   typedef int*			__c_locale;
++#endif
  
    // Convert numeric value of type double and long double to string and
    // return length of string.  If vsnprintf is available use it, otherwise
-Index: gcc-4.4.1/libstdc++-v3/config/os/gnu-linux/ctype_base.h
-===================================================================
---- gcc-4.4.1.orig/libstdc++-v3/config/os/gnu-linux/ctype_base.h	2009-08-06 23:42:27.402242608 -0700
-+++ gcc-4.4.1/libstdc++-v3/config/os/gnu-linux/ctype_base.h	2009-08-06 23:43:44.842241959 -0700
-@@ -33,14 +33,21 @@
-  */
- 
- // Information as gleaned from /usr/include/ctype.h
--  
-+
-+#include <features.h> 
-+#include <ctype.h> 
-+
- _GLIBCXX_BEGIN_NAMESPACE(std)
- 
-   /// @brief  Base class for ctype.
+diff --git a/libstdc++-v3/config/os/gnu-linux/ctype_base.h b/libstdc++-v3/config/os/gnu-linux/ctype_base.h
+index ff1f157..e88f569 100644
+--- a/libstdc++-v3/config/os/gnu-linux/ctype_base.h
++++ b/libstdc++-v3/config/os/gnu-linux/ctype_base.h
+@@ -40,7 +40,11 @@ _GLIBCXX_BEGIN_NAMESPACE(std)
    struct ctype_base
    {
      // Non-standard typedefs.
--    typedef const int* 		__to_type;
-+#ifdef __UCLIBC__ 
-+    typedef const __ctype_touplow_t*   __to_type; 
-+#else 
-+    typedef const int*         __to_type; 
++#ifdef __UCLIBC__
++    typedef const __ctype_touplow_t* __to_type;
++#else
+     typedef const int* 		__to_type;
 +#endif
  
      // NB: Offsets into ctype<char>::_M_table force a particular size
      // on the mask type. Because of this, we don't use an enum.
+diff --git a/libstdc++-v3/include/c/clocale b/libstdc++-v3/include/c/clocale
+index 7ea60f0..619f128 100644
+--- a/libstdc++-v3/include/c/clocale
++++ b/libstdc++-v3/include/c/clocale
+@@ -32,5 +32,8 @@
+ #pragma GCC system_header
+ 
+ #include_next <locale.h>
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
+ 
+ #endif
+diff --git a/libstdc++-v3/include/c_global/clocale b/libstdc++-v3/include/c_global/clocale
+index 7cf42e2..cd9b37a 100644
+--- a/libstdc++-v3/include/c_global/clocale
++++ b/libstdc++-v3/include/c_global/clocale
+@@ -42,6 +42,9 @@
+ 
+ #include <bits/c++config.h>
+ #include <locale.h>
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
+ 
+ #ifndef _GLIBCXX_CLOCALE
+ #define _GLIBCXX_CLOCALE 1
+diff --git a/libstdc++-v3/include/c_std/clocale b/libstdc++-v3/include/c_std/clocale
+index 36d8bd7..7905d5d 100644
+--- a/libstdc++-v3/include/c_std/clocale
++++ b/libstdc++-v3/include/c_std/clocale
+@@ -45,6 +45,10 @@
+ #include <bits/c++config.h>
+ #include <locale.h>
+ 
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
++
+ // Get rid of those macros defined in <locale.h> in lieu of real functions.
+ #undef setlocale
+ #undef localeconv
diff --git a/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch b/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch
new file mode 100644
index 0000000..d20e3d0
--- /dev/null
+++ b/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch
@@ -0,0 +1,255 @@
+on powerpc32 cmath has to be compiled with -mlong-double-128
+
+diff --git a/libstdc++-v3/include/tr1_impl/cmath b/libstdc++-v3/include/tr1_impl/cmath
+index 5807380..f395a0d 100644
+--- a/libstdc++-v3/include/tr1_impl/cmath
++++ b/libstdc++-v3/include/tr1_impl/cmath
+@@ -150,143 +150,213 @@ _GLIBCXX_BEGIN_NAMESPACE_TR1
+   // functions
+   using ::acosh;
+   using ::acoshf;
+-  using ::acoshl;
++#ifdef __LONG_DOUBLE_128__
++   using ::acoshl;
++#endif
+ 
+   using ::asinh;
+   using ::asinhf;
+-  using ::asinhl;
++#ifdef __LONG_DOUBLE_128__
++   using ::asinhl;
++#endif
+ 
+   using ::atanh;
+   using ::atanhf;
+-  using ::atanhl;
++#ifdef __LONG_DOUBLE_128__
++   using ::atanhl;
++#endif
+ 
+   using ::cbrt;
+   using ::cbrtf;
+-  using ::cbrtl;
++#ifdef __LONG_DOUBLE_128__
++   using ::cbrtl;
++#endif
+ 
+   using ::copysign;
+   using ::copysignf;
+-  using ::copysignl;
++#ifdef __LONG_DOUBLE_128__
++   using ::copysignl;
++#endif
+ 
+   using ::erf;
+   using ::erff;
+-  using ::erfl;
++#ifdef __LONG_DOUBLE_128__
++   using ::erfl;
++#endif
+ 
+   using ::erfc;
+   using ::erfcf;
+-  using ::erfcl;
++#ifdef __LONG_DOUBLE_128__
++   using ::erfcl;
++#endif
+ 
+   using ::exp2;
+   using ::exp2f;
+-  using ::exp2l;
++#ifdef __LONG_DOUBLE_128__
++   using ::exp2l;
++#endif
+ 
+   using ::expm1;
+   using ::expm1f;
+-  using ::expm1l;
++#ifdef __LONG_DOUBLE_128__
++   using ::expm1l;
++#endif
+ 
+   using ::fdim;
+   using ::fdimf;
+-  using ::fdiml;
++#ifdef __LONG_DOUBLE_128__
++   using ::fdiml;
++#endif
+ 
+   using ::fma;
+   using ::fmaf;
+-  using ::fmal;
++#ifdef __LONG_DOUBLE_128__
++   using ::fmal;
++#endif
+ 
+   using ::fmax;
+   using ::fmaxf;
+-  using ::fmaxl;
++#ifdef __LONG_DOUBLE_128__
++   using ::fmaxl;
++#endif
+ 
+   using ::fmin;
+   using ::fminf;
+-  using ::fminl;
++#ifdef __LONG_DOUBLE_128__
++   using ::fminl;
++#endif
+ 
+   using ::hypot;
+   using ::hypotf;
+-  using ::hypotl;
++#ifdef __LONG_DOUBLE_128__
++   using ::hypotl;
++#endif
+ 
+   using ::ilogb;
+   using ::ilogbf;
+-  using ::ilogbl;
++#ifdef __LONG_DOUBLE_128__
++   using ::ilogbl;
++#endif
+ 
+   using ::lgamma;
+   using ::lgammaf;
+-  using ::lgammal;
++#ifdef __LONG_DOUBLE_128__
++   using ::lgammal;
++#endif
+ 
+   using ::llrint;
+   using ::llrintf;
+-  using ::llrintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::llrintl;
++#endif
+ 
+   using ::llround;
+   using ::llroundf;
+-  using ::llroundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::llroundl;
++#endif
+ 
+   using ::log1p;
+   using ::log1pf;
+-  using ::log1pl;
++#ifdef __LONG_DOUBLE_128__
++   using ::log1pl;
++#endif
+ 
+   using ::log2;
+   using ::log2f;
+-  using ::log2l;
++#ifdef __LONG_DOUBLE_128__
++   using ::log2l;
++#endif
+ 
+   using ::logb;
+   using ::logbf;
+-  using ::logbl;
++#ifdef __LONG_DOUBLE_128__
++   using ::logbl;
++#endif
+ 
+   using ::lrint;
+   using ::lrintf;
+-  using ::lrintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::lrintl;
++#endif
+ 
+   using ::lround;
+   using ::lroundf;
+-  using ::lroundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::lroundl;
++#endif
+ 
+   using ::nan;
+   using ::nanf;
+-  using ::nanl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nanl;
++#endif
+ 
+   using ::nearbyint;
+   using ::nearbyintf;
+-  using ::nearbyintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nearbyintl;
++#endif
+ 
+   using ::nextafter;
+   using ::nextafterf;
+-  using ::nextafterl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nextafterl;
++#endif
+ 
+   using ::nexttoward;
+   using ::nexttowardf;
+-  using ::nexttowardl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nexttowardl;
++#endif
+ 
+   using ::remainder;
+   using ::remainderf;
+-  using ::remainderl;
++#ifdef __LONG_DOUBLE_128__
++   using ::remainderl;
++#endif
+ 
+   using ::remquo;
+   using ::remquof;
+-  using ::remquol;
++#ifdef __LONG_DOUBLE_128__
++   using ::remquol;
++#endif
+ 
+   using ::rint;
+   using ::rintf;
+-  using ::rintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::rintl;
++#endif
+ 
+   using ::round;
+   using ::roundf;
+-  using ::roundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::roundl;
++#endif
+ 
+   using ::scalbln;
+   using ::scalblnf;
+-  using ::scalblnl;
++#ifdef __LONG_DOUBLE_128__
++   using ::scalblnl;
++#endif
+ 
+   using ::scalbn;
+   using ::scalbnf;
+-  using ::scalbnl;
++#ifdef __LONG_DOUBLE_128__
++   using ::scalbnl;
++#endif
+ 
+   using ::tgamma;
+   using ::tgammaf;
+-  using ::tgammal;
++#ifdef __LONG_DOUBLE_128__
++   using ::tgammal;
++#endif
+ 
+   using ::trunc;
+   using ::truncf;
+-  using ::truncl;
++#ifdef __LONG_DOUBLE_128__
++   using ::truncl;
++#endif
+ 
+ #endif
+ 
-- 
1.7.1

Re: [PATCH][v2] gcc-cross: fix PV and make it compile

[Bernhard Reutner-Fischer]

On Fri, Jul 16, 2010 at 11:32:14PM +0200, Bernhard Reutner-Fischer wrote:

>index 2e24d63..40b46d6 100644
>--- a/recipes/gcc/gcc-svn.inc
>+++ b/recipes/gcc/gcc-svn.inc
>@@ -1,7 +1,7 @@
> DEFAULT_PREFERENCE = "-999"
> ARM_INSTRUCTION_SET = "arm"
> 
>-SRCREV = "158653"
>+SRCREV = "162223"
> 
> require gcc-common.inc
> 
>@@ -9,9 +9,9 @@ DEPENDS = "mpfr gmp libmpc"
> NATIVEDEPS = "mpfr-native gmp-native libmpc-native"
> 
> BINV = "4.6.0"
>-PV = "4.6+svnr${SRCPV}"
>+PKGV = "4.6+svnr${SRCPV}"
> 
>-FILESPATHPKG .= ":gcc-svn"
>+FILESPATHPKG .= ":gcc-${PV}"

If i don't munge PR then i will not rebuild if the revision changes.
With above we'll always end up with (e.g.) gcc-cross-svn-r0
which is rather inconvenient, no?

[PATCH][v3] gcc-cross: fix PV and make it compile

[Bernhard Reutner-Fischer]

- using PREFERRED_GCC_VERSION = "svn" didn't work since
  gcc-svn.inc messed with PV
- disable libstdcxx-pch for ppc
  pch1a_output would need to be compiled with -mlong-double-128
  to avoid
    libstdc++-v3/include/precompiled/stdc++.h -o \
	powerpc-oe-linux-uclibc/bits/stdc++.h.gch/O2g.gch
    [snip]
    libstdc++-v3/include/tr1_impl/cmath:153:11: error: '::acoshl' has not been declared
  install proper guards and furthermore disable libstdcxx-pch for now

Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
---
 recipes/gcc/gcc-svn.inc                            |   10 +-
 recipes/gcc/gcc-svn/gcc-flags-for-build.patch      |   26 --
 .../gcc-uclibc-locale-ctype_touplow_t.patch        |   99 +++++---
 .../powerpc32-c++-pch-long-double-128.patch        |  255 ++++++++++++++++++++
 4 files changed, 323 insertions(+), 67 deletions(-)
 create mode 100644 recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch

diff --git a/recipes/gcc/gcc-svn.inc b/recipes/gcc/gcc-svn.inc
index 2e24d63..233cd04 100644
--- a/recipes/gcc/gcc-svn.inc
+++ b/recipes/gcc/gcc-svn.inc
@@ -1,7 +1,7 @@
 DEFAULT_PREFERENCE = "-999"
 ARM_INSTRUCTION_SET = "arm"
 
-SRCREV = "158653"
+SRCREV = "162560"
 
 require gcc-common.inc
 
@@ -9,9 +9,9 @@ DEPENDS = "mpfr gmp libmpc"
 NATIVEDEPS = "mpfr-native gmp-native libmpc-native"
 
 BINV = "4.6.0"
-PV = "4.6+svnr${SRCPV}"
+PR = "4.6+svnr${SRCPV}"
 
-FILESPATHPKG .= ":gcc-svn"
+FILESPATHPKG .= ":gcc-${PV}"
 
 SRC_URI = "svn://gcc.gnu.org/svn/gcc;module=trunk \
 	   file://gcc-4.3.1-ARCH_FLAGS_FOR_TARGET.patch \
@@ -19,6 +19,7 @@ SRC_URI = "svn://gcc.gnu.org/svn/gcc;module=trunk \
            file://gcc-uclibc-locale-ctype_touplow_t.patch \
 	   file://cache-amnesia.patch \
 	   file://gcc-flags-for-build.patch \
+	   file://powerpc32-c++-pch-long-double-128.patch \
 	  "
 
 # Language Overrides
@@ -42,3 +43,6 @@ do_unpack_append() {
 EXTRA_OECONF_append_linux-uclibc = " --disable-decimal-float "
 EXTRA_OECONF_append_linux-uclibceabi = " --disable-decimal-float "
 EXTRA_OECONF_append_linux-uclibcspe = " --disable-decimal-float "
+# powerpc32 would need to build libstdc++-v3/include/cmath with -mlong-double-128
+# just disable PCH for ppc
+EXTRA_OECONF_append_powerpc-linux-uclibc = " --disable-libstdcxx-pch "
diff --git a/recipes/gcc/gcc-svn/gcc-flags-for-build.patch b/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
index 98b3ecf..0a0da9a 100644
--- a/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
+++ b/recipes/gcc/gcc-svn/gcc-flags-for-build.patch
@@ -98,24 +98,6 @@ Index: gcc-4.5+svnr155680/gcc/configure
  fi
  
  # Expand extra_headers to include complete path.
-@@ -17037,7 +17040,7 @@ else
-   lt_dlunknown=0; lt_dlno_uscore=1; lt_dlneed_uscore=2
-   lt_status=$lt_dlunknown
-   cat > conftest.$ac_ext <<_LT_EOF
--#line 17040 "configure"
-+#line 17043 "configure"
- #include "confdefs.h"
- 
- #if HAVE_DLFCN_H
-@@ -17143,7 +17146,7 @@ else
-   lt_dlunknown=0; lt_dlno_uscore=1; lt_dlneed_uscore=2
-   lt_status=$lt_dlunknown
-   cat > conftest.$ac_ext <<_LT_EOF
--#line 17146 "configure"
-+#line 17149 "configure"
- #include "confdefs.h"
- 
- #if HAVE_DLFCN_H
 Index: gcc-4.5+svnr155680/Makefile.tpl
 ===================================================================
 --- gcc-4.5+svnr155680.orig/Makefile.tpl	2010-01-13 01:02:22.603540396 -0800
@@ -148,14 +130,6 @@ Index: gcc-4.5+svnr155680/configure
    CXXFLAGS_FOR_BUILD=${CXXFLAGS_FOR_BUILD-${CXXFLAGS}}
    LDFLAGS_FOR_BUILD=${LDFLAGS_FOR_BUILD-${LDFLAGS}}
  fi
-@@ -7940,6 +7942,7 @@ done
- 
- 
- 
-+
- # Generate default definitions for YACC, M4, LEX and other programs that run
- # on the build machine.  These are used if the Makefile can't locate these
- # programs in objdir.
 Index: gcc-4.5+svnr155680/configure.ac
 ===================================================================
 --- gcc-4.5+svnr155680.orig/configure.ac	2010-01-13 01:02:22.651540130 -0800
diff --git a/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch b/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
index ffb37d2..9cae744 100644
--- a/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
+++ b/recipes/gcc/gcc-svn/gcc-uclibc-locale-ctype_touplow_t.patch
@@ -1,52 +1,75 @@
-Index: gcc-4.4.1/libstdc++-v3/config/locale/generic/c_locale.h
-===================================================================
---- gcc-4.4.1.orig/libstdc++-v3/config/locale/generic/c_locale.h	2009-08-06 23:38:32.398265633 -0700
-+++ gcc-4.4.1/libstdc++-v3/config/locale/generic/c_locale.h	2009-08-06 23:41:09.778242281 -0700
-@@ -41,12 +41,17 @@
- 
- #include <clocale>
- #include <cstddef>
-+#include <features.h> 
-+#include <ctype.h> 
- 
+diff --git a/libstdc++-v3/config/locale/generic/c_locale.h b/libstdc++-v3/config/locale/generic/c_locale.h
+index 7d685b8..4f69857 100644
+--- a/libstdc++-v3/config/locale/generic/c_locale.h
++++ b/libstdc++-v3/config/locale/generic/c_locale.h
+@@ -44,8 +44,11 @@
  #define _GLIBCXX_NUM_CATEGORIES 0
  
  _GLIBCXX_BEGIN_NAMESPACE(std)
 -
--  typedef int*			__c_locale;
-+#ifdef __UCLIBC__ 
-+   typedef __ctype_touplow_t*   __c_locale; 
-+#else 
-+   typedef int*         __c_locale; 
-+#endif 
++#ifdef __UCLIBC__
++  typedef __ctype_touplow_t* __c_locale;
++#else
+   typedef int*			__c_locale;
++#endif
  
    // Convert numeric value of type double and long double to string and
    // return length of string.  If vsnprintf is available use it, otherwise
-Index: gcc-4.4.1/libstdc++-v3/config/os/gnu-linux/ctype_base.h
-===================================================================
---- gcc-4.4.1.orig/libstdc++-v3/config/os/gnu-linux/ctype_base.h	2009-08-06 23:42:27.402242608 -0700
-+++ gcc-4.4.1/libstdc++-v3/config/os/gnu-linux/ctype_base.h	2009-08-06 23:43:44.842241959 -0700
-@@ -33,14 +33,21 @@
-  */
- 
- // Information as gleaned from /usr/include/ctype.h
--  
-+
-+#include <features.h> 
-+#include <ctype.h> 
-+
- _GLIBCXX_BEGIN_NAMESPACE(std)
- 
-   /// @brief  Base class for ctype.
+diff --git a/libstdc++-v3/config/os/gnu-linux/ctype_base.h b/libstdc++-v3/config/os/gnu-linux/ctype_base.h
+index ff1f157..e88f569 100644
+--- a/libstdc++-v3/config/os/gnu-linux/ctype_base.h
++++ b/libstdc++-v3/config/os/gnu-linux/ctype_base.h
+@@ -40,7 +40,11 @@ _GLIBCXX_BEGIN_NAMESPACE(std)
    struct ctype_base
    {
      // Non-standard typedefs.
--    typedef const int* 		__to_type;
-+#ifdef __UCLIBC__ 
-+    typedef const __ctype_touplow_t*   __to_type; 
-+#else 
-+    typedef const int*         __to_type; 
++#ifdef __UCLIBC__
++    typedef const __ctype_touplow_t* __to_type;
++#else
+     typedef const int* 		__to_type;
 +#endif
  
      // NB: Offsets into ctype<char>::_M_table force a particular size
      // on the mask type. Because of this, we don't use an enum.
+diff --git a/libstdc++-v3/include/c/clocale b/libstdc++-v3/include/c/clocale
+index 7ea60f0..619f128 100644
+--- a/libstdc++-v3/include/c/clocale
++++ b/libstdc++-v3/include/c/clocale
+@@ -32,5 +32,8 @@
+ #pragma GCC system_header
+ 
+ #include_next <locale.h>
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
+ 
+ #endif
+diff --git a/libstdc++-v3/include/c_global/clocale b/libstdc++-v3/include/c_global/clocale
+index 7cf42e2..cd9b37a 100644
+--- a/libstdc++-v3/include/c_global/clocale
++++ b/libstdc++-v3/include/c_global/clocale
+@@ -42,6 +42,9 @@
+ 
+ #include <bits/c++config.h>
+ #include <locale.h>
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
+ 
+ #ifndef _GLIBCXX_CLOCALE
+ #define _GLIBCXX_CLOCALE 1
+diff --git a/libstdc++-v3/include/c_std/clocale b/libstdc++-v3/include/c_std/clocale
+index 36d8bd7..7905d5d 100644
+--- a/libstdc++-v3/include/c_std/clocale
++++ b/libstdc++-v3/include/c_std/clocale
+@@ -45,6 +45,10 @@
+ #include <bits/c++config.h>
+ #include <locale.h>
+ 
++#ifdef __UCLIBC__
++#include <ctype.h>
++#endif
++
+ // Get rid of those macros defined in <locale.h> in lieu of real functions.
+ #undef setlocale
+ #undef localeconv
diff --git a/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch b/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch
new file mode 100644
index 0000000..d20e3d0
--- /dev/null
+++ b/recipes/gcc/gcc-svn/powerpc32-c++-pch-long-double-128.patch
@@ -0,0 +1,255 @@
+on powerpc32 cmath has to be compiled with -mlong-double-128
+
+diff --git a/libstdc++-v3/include/tr1_impl/cmath b/libstdc++-v3/include/tr1_impl/cmath
+index 5807380..f395a0d 100644
+--- a/libstdc++-v3/include/tr1_impl/cmath
++++ b/libstdc++-v3/include/tr1_impl/cmath
+@@ -150,143 +150,213 @@ _GLIBCXX_BEGIN_NAMESPACE_TR1
+   // functions
+   using ::acosh;
+   using ::acoshf;
+-  using ::acoshl;
++#ifdef __LONG_DOUBLE_128__
++   using ::acoshl;
++#endif
+ 
+   using ::asinh;
+   using ::asinhf;
+-  using ::asinhl;
++#ifdef __LONG_DOUBLE_128__
++   using ::asinhl;
++#endif
+ 
+   using ::atanh;
+   using ::atanhf;
+-  using ::atanhl;
++#ifdef __LONG_DOUBLE_128__
++   using ::atanhl;
++#endif
+ 
+   using ::cbrt;
+   using ::cbrtf;
+-  using ::cbrtl;
++#ifdef __LONG_DOUBLE_128__
++   using ::cbrtl;
++#endif
+ 
+   using ::copysign;
+   using ::copysignf;
+-  using ::copysignl;
++#ifdef __LONG_DOUBLE_128__
++   using ::copysignl;
++#endif
+ 
+   using ::erf;
+   using ::erff;
+-  using ::erfl;
++#ifdef __LONG_DOUBLE_128__
++   using ::erfl;
++#endif
+ 
+   using ::erfc;
+   using ::erfcf;
+-  using ::erfcl;
++#ifdef __LONG_DOUBLE_128__
++   using ::erfcl;
++#endif
+ 
+   using ::exp2;
+   using ::exp2f;
+-  using ::exp2l;
++#ifdef __LONG_DOUBLE_128__
++   using ::exp2l;
++#endif
+ 
+   using ::expm1;
+   using ::expm1f;
+-  using ::expm1l;
++#ifdef __LONG_DOUBLE_128__
++   using ::expm1l;
++#endif
+ 
+   using ::fdim;
+   using ::fdimf;
+-  using ::fdiml;
++#ifdef __LONG_DOUBLE_128__
++   using ::fdiml;
++#endif
+ 
+   using ::fma;
+   using ::fmaf;
+-  using ::fmal;
++#ifdef __LONG_DOUBLE_128__
++   using ::fmal;
++#endif
+ 
+   using ::fmax;
+   using ::fmaxf;
+-  using ::fmaxl;
++#ifdef __LONG_DOUBLE_128__
++   using ::fmaxl;
++#endif
+ 
+   using ::fmin;
+   using ::fminf;
+-  using ::fminl;
++#ifdef __LONG_DOUBLE_128__
++   using ::fminl;
++#endif
+ 
+   using ::hypot;
+   using ::hypotf;
+-  using ::hypotl;
++#ifdef __LONG_DOUBLE_128__
++   using ::hypotl;
++#endif
+ 
+   using ::ilogb;
+   using ::ilogbf;
+-  using ::ilogbl;
++#ifdef __LONG_DOUBLE_128__
++   using ::ilogbl;
++#endif
+ 
+   using ::lgamma;
+   using ::lgammaf;
+-  using ::lgammal;
++#ifdef __LONG_DOUBLE_128__
++   using ::lgammal;
++#endif
+ 
+   using ::llrint;
+   using ::llrintf;
+-  using ::llrintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::llrintl;
++#endif
+ 
+   using ::llround;
+   using ::llroundf;
+-  using ::llroundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::llroundl;
++#endif
+ 
+   using ::log1p;
+   using ::log1pf;
+-  using ::log1pl;
++#ifdef __LONG_DOUBLE_128__
++   using ::log1pl;
++#endif
+ 
+   using ::log2;
+   using ::log2f;
+-  using ::log2l;
++#ifdef __LONG_DOUBLE_128__
++   using ::log2l;
++#endif
+ 
+   using ::logb;
+   using ::logbf;
+-  using ::logbl;
++#ifdef __LONG_DOUBLE_128__
++   using ::logbl;
++#endif
+ 
+   using ::lrint;
+   using ::lrintf;
+-  using ::lrintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::lrintl;
++#endif
+ 
+   using ::lround;
+   using ::lroundf;
+-  using ::lroundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::lroundl;
++#endif
+ 
+   using ::nan;
+   using ::nanf;
+-  using ::nanl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nanl;
++#endif
+ 
+   using ::nearbyint;
+   using ::nearbyintf;
+-  using ::nearbyintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nearbyintl;
++#endif
+ 
+   using ::nextafter;
+   using ::nextafterf;
+-  using ::nextafterl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nextafterl;
++#endif
+ 
+   using ::nexttoward;
+   using ::nexttowardf;
+-  using ::nexttowardl;
++#ifdef __LONG_DOUBLE_128__
++   using ::nexttowardl;
++#endif
+ 
+   using ::remainder;
+   using ::remainderf;
+-  using ::remainderl;
++#ifdef __LONG_DOUBLE_128__
++   using ::remainderl;
++#endif
+ 
+   using ::remquo;
+   using ::remquof;
+-  using ::remquol;
++#ifdef __LONG_DOUBLE_128__
++   using ::remquol;
++#endif
+ 
+   using ::rint;
+   using ::rintf;
+-  using ::rintl;
++#ifdef __LONG_DOUBLE_128__
++   using ::rintl;
++#endif
+ 
+   using ::round;
+   using ::roundf;
+-  using ::roundl;
++#ifdef __LONG_DOUBLE_128__
++   using ::roundl;
++#endif
+ 
+   using ::scalbln;
+   using ::scalblnf;
+-  using ::scalblnl;
++#ifdef __LONG_DOUBLE_128__
++   using ::scalblnl;
++#endif
+ 
+   using ::scalbn;
+   using ::scalbnf;
+-  using ::scalbnl;
++#ifdef __LONG_DOUBLE_128__
++   using ::scalbnl;
++#endif
+ 
+   using ::tgamma;
+   using ::tgammaf;
+-  using ::tgammal;
++#ifdef __LONG_DOUBLE_128__
++   using ::tgammal;
++#endif
+ 
+   using ::trunc;
+   using ::truncf;
+-  using ::truncl;
++#ifdef __LONG_DOUBLE_128__
++   using ::truncl;
++#endif
+ 
+ #endif
+ 
-- 
1.7.1

Re: [PATCH][v3] gcc-cross: fix PV and make it compile

[Koen Kooi]

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

On 28-07-10 19:14, Bernhard Reutner-Fischer wrote:

> -PV = "4.6+svnr${SRCPV}"
> +PR = "4.6+svnr${SRCPV}"

NAK, PV = "4.6+svnr${SRCPV}" is the right thing to do
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFMUGXVMkyGM64RGpERAvafAJ9g8mmWheRTk+uh1OL4k0kVSp6tNQCcDYQL
taWSWi0X3fXIEbe6x/A8XFI=
=GLUK
-----END PGP SIGNATURE-----

Re: [PATCH][v3] gcc-cross: fix PV and make it compile

[Bernhard Reutner-Fischer]

On Wed, Jul 28, 2010 at 07:16:05PM +0200, Koen Kooi wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>On 28-07-10 19:14, Bernhard Reutner-Fischer wrote:
>
>> -PV = "4.6+svnr${SRCPV}"
>> +PR = "4.6+svnr${SRCPV}"
>
>NAK, PV = "4.6+svnr${SRCPV}" is the right thing to do

ok what should i do then?
I want to use the svn version of gcc regardless of it being called
"1.2.3-svn" or "5.0.0-svn" so i put this into my local.conf:

#PREFERRED_GCC_VERSION = "4.4.4"
PREFERRED_GCC_VERSION = "svn"

But that obviously doesn't work with a PV of "4.6+svnr${SRCPV}" like
current git has.
I made up that "svn" intuitively, with the following reasoning:
i'm using the subversion version (_version_, hence PV)
with an arbitrary revision as Package Release that will sort correctly
iff a 4.6.0 release would be made.

Now how am i supposed to pick "svn" if not by using "svn"?

thanks for clarification,

Re: [PATCH][v3] gcc-cross: fix PV and make it compile

[Koen Kooi]

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


http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-July/021641.html

On 28-07-10 19:39, Bernhard Reutner-Fischer wrote:
> On Wed, Jul 28, 2010 at 07:16:05PM +0200, Koen Kooi wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 28-07-10 19:14, Bernhard Reutner-Fischer wrote:
>>
>>> -PV = "4.6+svnr${SRCPV}"
>>> +PR = "4.6+svnr${SRCPV}"
>>
>> NAK, PV = "4.6+svnr${SRCPV}" is the right thing to do
> 
> ok what should i do then?
> I want to use the svn version of gcc regardless of it being called
> "1.2.3-svn" or "5.0.0-svn" so i put this into my local.conf:
> 
> #PREFERRED_GCC_VERSION = "4.4.4"
> PREFERRED_GCC_VERSION = "svn"
> 
> But that obviously doesn't work with a PV of "4.6+svnr${SRCPV}" like
> current git has.
> I made up that "svn" intuitively, with the following reasoning:
> i'm using the subversion version (_version_, hence PV)
> with an arbitrary revision as Package Release that will sort correctly
> iff a 4.6.0 release would be made.
> 
> Now how am i supposed to pick "svn" if not by using "svn"?
> 
> thanks for clarification,

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

iD8DBQFMUIbfMkyGM64RGpERAoPYAJ9rxqMp47aGQnPjE6R7GTIMx7CQpgCgmuIa
7su2ka+K0cQBaopKy11SRtk=
=1Jf8
-----END PGP SIGNATURE-----

Re: [PATCH][v3] gcc-cross: fix PV and make it compile

[Bernhard Reutner-Fischer]

On Wed, Jul 28, 2010 at 09:37:03PM +0200, Koen Kooi wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>
>http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-July/021641.html

this is rather ugly and counter intuitive for i'll have to change the
PREFERRED_GCC_VERSION whenever someone decides to change that "4.6+svnr",
but ok, i'll drop the PR hunk :(
thanks,

>On 28-07-10 19:39, Bernhard Reutner-Fischer wrote:
>> On Wed, Jul 28, 2010 at 07:16:05PM +0200, Koen Kooi wrote:
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>> On 28-07-10 19:14, Bernhard Reutner-Fischer wrote:
>>>
>>>> -PV = "4.6+svnr${SRCPV}"
>>>> +PR = "4.6+svnr${SRCPV}"
>>>
>>> NAK, PV = "4.6+svnr${SRCPV}" is the right thing to do
>> 
>> ok what should i do then?
>> I want to use the svn version of gcc regardless of it being called
>> "1.2.3-svn" or "5.0.0-svn" so i put this into my local.conf:
>> 
>> #PREFERRED_GCC_VERSION = "4.4.4"
>> PREFERRED_GCC_VERSION = "svn"
>> 
>> But that obviously doesn't work with a PV of "4.6+svnr${SRCPV}" like
>> current git has.
>> I made up that "svn" intuitively, with the following reasoning:
>> i'm using the subversion version (_version_, hence PV)
>> with an arbitrary revision as Package Release that will sort correctly
>> iff a 4.6.0 release would be made.
>> 
>> Now how am i supposed to pick "svn" if not by using "svn"?
>> 
>> thanks for clarification,
>
>-----BEGIN PGP SIGNATURE-----
>Version: GnuPG v1.4.5 (Darwin)
>
>iD8DBQFMUIbfMkyGM64RGpERAoPYAJ9rxqMp47aGQnPjE6R7GTIMx7CQpgCgmuIa
>7su2ka+K0cQBaopKy11SRtk=
>=1Jf8
>-----END PGP SIGNATURE-----
>
>
>_______________________________________________
>Openembedded-devel mailing list
>Openembedded-devel@lists.openembedded.org
>http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

Re: [PATCH][v3] gcc-cross: fix PV and make it compile

[Frans Meulenbroeks]

2010/7/28 Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>

> On Wed, Jul 28, 2010 at 09:37:03PM +0200, Koen Kooi wrote:
> >-----BEGIN PGP SIGNED MESSAGE-----
> >Hash: SHA1
> >
> >
> >
> http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-July/021641.html
>
> this is rather ugly and counter intuitive for i'll have to change the
> PREFERRED_GCC_VERSION whenever someone decides to change that "4.6+svnr",
> but ok, i'll drop the PR hunk :(
> thanks,
>

Alternate suggestion: do it the way you want to do it and store the recipe
in a local overlay
You might also be able to use amend.inc.

Have fun! Frans

>
> >On 28-07-10 19:39, Bernhard Reutner-Fischer wrote:
> >> On Wed, Jul 28, 2010 at 07:16:05PM +0200, Koen Kooi wrote:
> >>> -----BEGIN PGP SIGNED MESSAGE-----
> >>> Hash: SHA1
> >>>
> >>> On 28-07-10 19:14, Bernhard Reutner-Fischer wrote:
> >>>
> >>>> -PV = "4.6+svnr${SRCPV}"
> >>>> +PR = "4.6+svnr${SRCPV}"
> >>>
> >>> NAK, PV = "4.6+svnr${SRCPV}" is the right thing to do
> >>
> >> ok what should i do then?
> >> I want to use the svn version of gcc regardless of it being called
> >> "1.2.3-svn" or "5.0.0-svn" so i put this into my local.conf:
> >>
> >> #PREFERRED_GCC_VERSION = "4.4.4"
> >> PREFERRED_GCC_VERSION = "svn"
> >>
> >> But that obviously doesn't work with a PV of "4.6+svnr${SRCPV}" like
> >> current git has.
> >> I made up that "svn" intuitively, with the following reasoning:
> >> i'm using the subversion version (_version_, hence PV)
> >> with an arbitrary revision as Package Release that will sort correctly
> >> iff a 4.6.0 release would be made.
> >>
> >> Now how am i supposed to pick "svn" if not by using "svn"?
> >>
> >> thanks for clarification,
> >
> >-----BEGIN PGP SIGNATURE-----
> >Version: GnuPG v1.4.5 (Darwin)
> >
> >iD8DBQFMUIbfMkyGM64RGpERAoPYAJ9rxqMp47aGQnPjE6R7GTIMx7CQpgCgmuIa
> >7su2ka+K0cQBaopKy11SRtk=
> >=1Jf8
> >-----END PGP SIGNATURE-----
> >
> >
> >_______________________________________________
> >Openembedded-devel mailing list
> >Openembedded-devel@lists.openembedded.org
> >http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>