Re: devel/boost: new patch for testing update to 1.39

2009-08-04 Thread Alexander Churanov
2009/8/4 Max Brazhnikov :
> there is no any port depending on pyste and I doubt there will be one in near
> future. Pulling gcc3 for newer systems because of hardly potential
> inconvenience seems to be excessive for me.

OK, if there are really no ports depending on it, then I'll implement
the proposed WITH_PYSTE switch, without mangling it in the package
name. Users who specifycally need Pyste will have to build their
custom version of boost-python-libs and deinstall/reinstall.

Sincerely,
Alexander Churanov,
maintainer of devel/boost-*
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"


Re: devel/boost: new patch for testing update to 1.39

2009-08-04 Thread Max Brazhnikov
On Tue, 4 Aug 2009 14:28:02 +0400, Alexander Churanov wrote:
> 2009/8/4 Mel Flynn :
> > I think for the time being PySTE support should be off by default,
> > because it pulls in a second libc and databases/akonadi plus kdeedu[34]
> > use boost-python and because gccxml looks like a dead project.
>
> Agree.
PySTE was disabled before boost split, would be nice to have it disabled for 
now also.

> > ports/123927 could be solved/prevented by using a WANT_BOOST_PYSTE in
> > dependent ports that rely on boost having PYSTE and using PKGNAMESUFFIX
> > plus pkg_info to verify if boost-python-libs is boost-python-pyste-libs -
> > as a stop-gap.
>
> Well, if the user installs some port that depends on
> boost-python-libs, the user gets boost-python-libs installed. Then the
> user wants to install another port that depends on
> boost-python-pyste-libs. Would this succeed?

there is no any port depending on pyste and I doubt there will be one in near 
future. Pulling gcc3 for newer systems because of hardly potential 
inconvenience seems to be excessive for me.

Regards,
Max


signature.asc
Description: This is a digitally signed message part.


Re: devel/boost: new patch for testing update to 1.39

2009-08-04 Thread Alexander Churanov
2009/8/4 Mel Flynn :
>
> I think for the time being PySTE support should be off by default, because it
> pulls in a second libc and databases/akonadi plus kdeedu[34] use boost-python
> and because gccxml looks like a dead project.

Agree.

>
> I've attached a patch that does exactly this and also fixes WITH_DEBUG
> building. There's one issue commented in the Makefile that I'm unable to
> relate to anything I did, but am assuming it's from WITH_DEBUG as well.

The issue was I've accidently hardcoded path which built python
libraries are copied from. As a result, installation success was
depending not only on DEBUG setting, but also on the version of gcc
and, probably, other things.

This is already fixed and checked-in to CVS.

>
> ports/123927 could be solved/prevented by using a WANT_BOOST_PYSTE in
> dependent ports that rely on boost having PYSTE and using PKGNAMESUFFIX plus
> pkg_info to verify if boost-python-libs is boost-python-pyste-libs - as a
> stop-gap.

Well, if the user installs some port that depends on
boost-python-libs, the user gets boost-python-libs installed. Then the
user wants to install another port that depends on
boost-python-pyste-libs. Would this succeed?

Sincerely,
Alexander Churanov,
maintainer of devel/boost-*
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"


Re: [kde-freebsd] devel/boost: new patch for testing update to 1.39

2009-08-04 Thread Martin Wilke
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Mon, Aug 03, 2009 at 01:53:55PM -0800, Mel Flynn wrote:
> On Friday 31 July 2009 04:29:06 Alexander Churanov wrote:
> 
> > Now It's clear. Thank you for explanation.
> > The WITH_PYSTE was not in the options framework, that's why I've
> > missed this part of work.
> >
> > I see two possible solutions:
> >
> > 1) Put the WITH_PYSTE variable back.
> > 2) Create a separate port like devel/boost-pyste.
> >
> > The first is easy, however leads to dependeny issues like 'let's
> > verify that  boost-python-libs is installed with PYSTE support'.
> > Provided that Pyste is actually used by someone, this may re-introduce
> > bug ports/123927 (see
> > http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/123927).
> >
> > The second is much more robust, but requires more work. This is also
> > harder because I am not a user of Boost.Pyste. Thus, correctness
> > verification would be expensive.
> 
> I think for the time being PySTE support should be off by default, because it 
> pulls in a second libc and databases/akonadi plus kdeedu[34] use boost-python 
> and because gccxml looks like a dead project.
> 
> I've attached a patch that does exactly this and also fixes WITH_DEBUG 
> building. There's one issue commented in the Makefile that I'm unable to 
> relate to anything I did, but am assuming it's from WITH_DEBUG as well.
> 
> ports/123927 could be solved/prevented by using a WANT_BOOST_PYSTE in 
> dependent ports that rely on boost having PYSTE and using PKGNAMESUFFIX plus 
> pkg_info to verify if boost-python-libs is boost-python-pyste-libs - as a 
> stop-gap.

+1 

> 
> -- 
> Mel

> --- Makefile.orig 2009-07-28 03:44:50.0 -0800
> +++ Makefile  2009-08-03 00:55:34.0 -0800
> @@ -25,9 +25,23 @@
>   PYTHON_INCLUDES="${PYTHON_INCLUDEDIR}"\
>   PYTHON_LIB_PATH="${PYTHON_LIBDIR}"
>  
> -RUN_DEPENDS+=
> ${PREFIX}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
> -RUN_DEPENDS+=gccxml:${PORTSDIR}/devel/gccxml \
> - 
> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
> +RUN_DEPENDS+=
> ${LOCALBASE}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
> +RUN_DEPENDS+=
> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
> +.if defined(WITH_PYSTE)
> +RUN_DEPENDS+=gccxml:${PORTSDIR}/devel/gccxml
> +PLIST_SUB+=  PYSTE=
> +.else
> +PLIST_SUB+=  PYSTE="@comment "
> +.endif
> +
> +.if defined(WITH_DEBUG)
> +BOOST_BUILD= debug
> +# No idea how this gets enabled, but I'm getting it
> +OPTIM_SPEED= optimization-speed/
> +.else
> +OPTIM_SPEED= /
> +BOOST_BUILD= release
> +.endif
>  
>  BUILD_DEPENDS+=  ${BJAM}:${PORTSDIR}/devel/boost-jam
>  BUILD_DEPENDS+=  
> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
> @@ -39,12 +53,13 @@
>  
>  post-patch: customize-boost-build
>  
> -
> +.if defined(WITH_PYSTE)
>  post-configure:
>  # Configure pyste, the Boost.Python code generator
>   @cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
>   ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>   config ${PYDISTUTILS_CONFIGUREARGS}
> +.endif
>  
>  do-build:
>  # build the library
> @@ -52,19 +67,20 @@
>   ${SETENV} ${MAKE_ENV} ${BJAM} ${BJAM_OPTIONS} 
> --prefix=${PREFIX}\
>   --toolset=${BOOST_TOOLS} --with-python
>  
> +.if defined(WITH_PYSTE)
>  # Build pyste, the Boost.Python code generator
>   cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
>   ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>   build ${PYDISTUTILS_BUILDARGS}
> -
> +.endif
>  # Remove non-python headers
>   cd ${WRKSRC} && \
>   ${FIND} boost -not -path '*python*' -delete
>  
>  do-install:
>  # Install libraries
> - ${INSTALL_DATA} 
> ${WRKSRC}/bin.v2/libs/python/build/gcc-4.2.1/release/link-static/threading-multi/libboost_python.a
>  ${PREFIX}/lib
> - ${INSTALL_DATA} 
> ${WRKSRC}/bin.v2/libs/python/build/gcc-4.2.1/release/threading-multi/libboost_python.so.${PORTVERSION}
>  ${PREFIX}/lib/libboost_python.so.${BOOST_SHARED_LIB_VER}
> + ${INSTALL_DATA} 
> ${WRKSRC}/bin.v2/libs/python/build/gcc-4.2.1/${BOOST_BUILD}/link-static/${OPTIM_SPEED}threading-multi/libboost_python.a
>  ${PREFIX}/lib
> + ${INSTALL_DATA} 
> ${WRKSRC}/bin.v2/libs/python/build/gcc-4.2.1/${BOOST_BUILD}/${OPTIM_SPEED}threading-multi/libboost_python.so.${PORTVERSION}
>  ${PREFIX}/lib/libboost_python.so.${BOOST_SHARED_LIB_VER}
>  
>  # Create symlink for shared library
>   cd ${PREFIX}/lib/ && \
> @@ -79,10 +95,12 @@
>   ${FIND} boost -type f -a \! -name "*.orig"\
>   -exec ${INSTALL_DATA} \{\} ${PREFIX}/include/\{\} \;
>  
> +.if defined(WITH_PYSTE)
>  # Install pyste, the Boost.Python code generator
>   cd ${WRKSRC}/libs/python/pyste/install && \
>   ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>

Re: devel/boost: new patch for testing update to 1.39

2009-08-03 Thread Mel Flynn
On Friday 31 July 2009 04:29:06 Alexander Churanov wrote:

> Now It's clear. Thank you for explanation.
> The WITH_PYSTE was not in the options framework, that's why I've
> missed this part of work.
>
> I see two possible solutions:
>
> 1) Put the WITH_PYSTE variable back.
> 2) Create a separate port like devel/boost-pyste.
>
> The first is easy, however leads to dependeny issues like 'let's
> verify that  boost-python-libs is installed with PYSTE support'.
> Provided that Pyste is actually used by someone, this may re-introduce
> bug ports/123927 (see
> http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/123927).
>
> The second is much more robust, but requires more work. This is also
> harder because I am not a user of Boost.Pyste. Thus, correctness
> verification would be expensive.

I think for the time being PySTE support should be off by default, because it 
pulls in a second libc and databases/akonadi plus kdeedu[34] use boost-python 
and because gccxml looks like a dead project.

I've attached a patch that does exactly this and also fixes WITH_DEBUG 
building. There's one issue commented in the Makefile that I'm unable to 
relate to anything I did, but am assuming it's from WITH_DEBUG as well.

ports/123927 could be solved/prevented by using a WANT_BOOST_PYSTE in 
dependent ports that rely on boost having PYSTE and using PKGNAMESUFFIX plus 
pkg_info to verify if boost-python-libs is boost-python-pyste-libs - as a 
stop-gap.

-- 
Mel
--- Makefile.orig	2009-07-28 03:44:50.0 -0800
+++ Makefile	2009-08-03 00:55:34.0 -0800
@@ -25,9 +25,23 @@
 		PYTHON_INCLUDES="${PYTHON_INCLUDEDIR}"\
 		PYTHON_LIB_PATH="${PYTHON_LIBDIR}"
 
-RUN_DEPENDS+=	${PREFIX}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
-RUN_DEPENDS+=	gccxml:${PORTSDIR}/devel/gccxml \
-		${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
+RUN_DEPENDS+=	${LOCALBASE}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
+RUN_DEPENDS+=	${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
+.if defined(WITH_PYSTE)
+RUN_DEPENDS+=	gccxml:${PORTSDIR}/devel/gccxml
+PLIST_SUB+=	PYSTE=
+.else
+PLIST_SUB+=	PYSTE="@comment "
+.endif
+
+.if defined(WITH_DEBUG)
+BOOST_BUILD=	debug
+# No idea how this gets enabled, but I'm getting it
+OPTIM_SPEED=	optimization-speed/
+.else
+OPTIM_SPEED=	/
+BOOST_BUILD=	release
+.endif
 
 BUILD_DEPENDS+=	${BJAM}:${PORTSDIR}/devel/boost-jam
 BUILD_DEPENDS+=	${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
@@ -39,12 +53,13 @@
 
 post-patch: customize-boost-build
 
-
+.if defined(WITH_PYSTE)
 post-configure:
 # Configure pyste, the Boost.Python code generator
 	@cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
 	${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
 		config ${PYDISTUTILS_CONFIGUREARGS}
+.endif
 
 do-build:
 # build the library
@@ -52,19 +67,20 @@
 		${SETENV} ${MAKE_ENV} ${BJAM} ${BJAM_OPTIONS} --prefix=${PREFIX}\
 		--toolset=${BOOST_TOOLS} --with-python
 
+.if defined(WITH_PYSTE)
 # Build pyste, the Boost.Python code generator
 	cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
 	${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
 		build ${PYDISTUTILS_BUILDARGS}
-
+.endif
 # Remove non-python headers
 	cd ${WRKSRC} && \
 	${FIND} boost -not -path '*python*' -delete
 
 do-install:
 # Install libraries
-	${INSTALL_DATA} ${WRKSRC}/bin.v2/libs/python/build/gcc-4.2.1/release/link-static/threading-multi/libboost_python.a ${PREFIX}/lib
-	${INSTALL_DATA} ${WRKSRC}/bin.v2/libs/python/build/gcc-4.2.1/release/threading-multi/libboost_python.so.${PORTVERSION} ${PREFIX}/lib/libboost_python.so.${BOOST_SHARED_LIB_VER}
+	${INSTALL_DATA} ${WRKSRC}/bin.v2/libs/python/build/gcc-4.2.1/${BOOST_BUILD}/link-static/${OPTIM_SPEED}threading-multi/libboost_python.a ${PREFIX}/lib
+	${INSTALL_DATA} ${WRKSRC}/bin.v2/libs/python/build/gcc-4.2.1/${BOOST_BUILD}/${OPTIM_SPEED}threading-multi/libboost_python.so.${PORTVERSION} ${PREFIX}/lib/libboost_python.so.${BOOST_SHARED_LIB_VER}
 
 # Create symlink for shared library
 	cd ${PREFIX}/lib/ && \
@@ -79,10 +95,12 @@
 	${FIND} boost -type f -a \! -name "*.orig"\
 		-exec ${INSTALL_DATA} \{\} ${PREFIX}/include/\{\} \;
 
+.if defined(WITH_PYSTE)
 # Install pyste, the Boost.Python code generator
 	cd ${WRKSRC}/libs/python/pyste/install && \
 	${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
 		install ${PYDISTUTILS_INSTALLARGS}
+.endif
 
 post-install:
 # display pkg-message
--- pkg-plist.orig	2009-07-28 03:44:50.0 -0800
+++ pkg-plist	2009-08-02 12:29:56.0 -0800
@@ -1,4 +1,4 @@
-bin/pyste.py
+%%PYSTE%%bin/pyste.py
 include/boost/python.hpp
 include/boost/python/arg_from_python.hpp
 include/boost/python/args.hpp
@@ -213,71 +213,71 @@
 lib/libboost_python.a
 lib/libboost_python.so
 lib/libboost_python.so.%%BOOST_SHARED_LIB_VER%%
-%%PYTHON_SITELIBDIR%%/Pyste-0.9.10-py%%PYTHON_VERSION%%.egg-info
-%%PYTHON_SITELIBDIR%%/Pyste/ClassExporter.py
-%%

Re: devel/boost: new patch for testing update to 1.39

2009-08-03 Thread Alexander Churanov
Folks,

I've try to prepeare boost-python-libs and boost-pyste split. However,
I can not promise any specific deadline. Perhaps, it will be done
within a month.

Sincerely,
Alexander Churanov,
maintainer of devel/boost-*


2009/7/31 Alexander Churanov :
> Mel,
>
> Now It's clear. Thank you for explanation.
> The WITH_PYSTE was not in the options framework, that's why I've
> missed this part of work.
>
> I see two possible solutions:
>
> 1) Put the WITH_PYSTE variable back.
> 2) Create a separate port like devel/boost-pyste.
>
> The first is easy, however leads to dependeny issues like 'let's
> verify that  boost-python-libs is installed with PYSTE support'.
> Provided that Pyste is actually used by someone, this may re-introduce
> bug ports/123927 (see
> http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/123927).
>
> The second is much more robust, but requires more work. This is also
> harder because I am not a user of Boost.Pyste. Thus, correctness
> verification would be expensive.
>
> Folks, is there a user of Boost.Pyste, willing to volunteer with
> devel/boost-pyste quality verification?
>
> Sincerely,
> Alexander Churanov
> maintainer of devel/boost-*
>
>
> 2009/7/30 Mel Flynn :
>> On Thursday 30 July 2009 03:20:56 Alexander Churanov wrote:
>>
>>> I do not understand.
>>>
>>> Is it correct that you are discussing the issue of Pyste being part of
>>> boost and depending on gccxml port, which has issues of its own?
>>>
>>> I do not use Pyste and did not performed any specific testing of it,
>>> but all of boost-python-libs builds and installs, including
>>> dependencies.
>>>
>>> Please, describe what do you expect from ports and what do you
>>> observe, what negative outcome the issue has.
>>
>> The previous boost-python port had an *optional* dependency on pyste.
>> The current one does not. This means that boost-python now pulls in
>> gccxml which pulls in lang/gcc34 on anything but 6.x.
>> I'm currently building boost-python-libs with the pyste stuff
>> ifdef'd out[1] and will try to build packages kdeedu3, kdebase4,
>> games/wesnoth to see if there's any adverse side-effects.
>>
>> --
>> Mel
>>
>> [1]
>> --- Makefile.orig       2009-07-28 03:44:50.0 -0800
>> +++ Makefile    2009-07-30 07:19:26.0 -0800
>> @@ -25,9 +25,11 @@
>>                PYTHON_INCLUDES="${PYTHON_INCLUDEDIR}"\
>>                PYTHON_LIB_PATH="${PYTHON_LIBDIR}"
>>
>> -RUN_DEPENDS+=  
>> ${PREFIX}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
>> -RUN_DEPENDS+=  gccxml:${PORTSDIR}/devel/gccxml \
>> -               
>> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
>> +RUN_DEPENDS+=  
>> ${LOCALBASE}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
>> +RUN_DEPENDS+=  
>> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
>> +.if defined(WITH_PYSTE)
>> +RUN_DEPENDS+=  gccxml:${PORTSDIR}/devel/gccxml
>> +.endif
>>
>>  BUILD_DEPENDS+=        ${BJAM}:${PORTSDIR}/devel/boost-jam
>>  BUILD_DEPENDS+=        
>> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
>> @@ -39,12 +41,13 @@
>>
>>  post-patch: customize-boost-build
>>
>> -
>> +.if defined(WITH_PYSTE)
>>  post-configure:
>>  # Configure pyste, the Boost.Python code generator
>>       �...@cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
>>        ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>>                config ${PYDISTUTILS_CONFIGUREARGS}
>> +.endif
>>
>>  do-build:
>>  # build the library
>> @@ -52,11 +55,12 @@
>>                ${SETENV} ${MAKE_ENV} ${BJAM} ${BJAM_OPTIONS} 
>> --prefix=${PREFIX}\
>>                --toolset=${BOOST_TOOLS} --with-python
>>
>> +.if defined(WITH_PYSTE)
>>  # Build pyste, the Boost.Python code generator
>>        cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
>>        ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>>                build ${PYDISTUTILS_BUILDARGS}
>> -
>> +.endif
>>  # Remove non-python headers
>>        cd ${WRKSRC} && \
>>        ${FIND} boost -not -path '*python*' -delete
>> @@ -79,10 +83,12 @@
>>        ${FIND} boost -type f -a \! -name "*.orig"\
>>                -exec ${INSTALL_DATA} \{\} ${PREFIX}/include/\{\} \;
>>
>> +.if defined(WITH_PYSTE)
>>  # Install pyste, the Boost.Python code generator
>>        cd ${WRKSRC}/libs/python/pyste/install && \
>>        ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>>                install ${PYDISTUTILS_INSTALLARGS}
>> +.endif
>>
>>  post-install:
>>  # display pkg-message
>>
>>
>
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"


Re: devel/boost: new patch for testing update to 1.39

2009-07-31 Thread Alexander Churanov
Mel,

Now It's clear. Thank you for explanation.
The WITH_PYSTE was not in the options framework, that's why I've
missed this part of work.

I see two possible solutions:

1) Put the WITH_PYSTE variable back.
2) Create a separate port like devel/boost-pyste.

The first is easy, however leads to dependeny issues like 'let's
verify that  boost-python-libs is installed with PYSTE support'.
Provided that Pyste is actually used by someone, this may re-introduce
bug ports/123927 (see
http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/123927).

The second is much more robust, but requires more work. This is also
harder because I am not a user of Boost.Pyste. Thus, correctness
verification would be expensive.

Folks, is there a user of Boost.Pyste, willing to volunteer with
devel/boost-pyste quality verification?

Sincerely,
Alexander Churanov
maintainer of devel/boost-*


2009/7/30 Mel Flynn :
> On Thursday 30 July 2009 03:20:56 Alexander Churanov wrote:
>
>> I do not understand.
>>
>> Is it correct that you are discussing the issue of Pyste being part of
>> boost and depending on gccxml port, which has issues of its own?
>>
>> I do not use Pyste and did not performed any specific testing of it,
>> but all of boost-python-libs builds and installs, including
>> dependencies.
>>
>> Please, describe what do you expect from ports and what do you
>> observe, what negative outcome the issue has.
>
> The previous boost-python port had an *optional* dependency on pyste.
> The current one does not. This means that boost-python now pulls in
> gccxml which pulls in lang/gcc34 on anything but 6.x.
> I'm currently building boost-python-libs with the pyste stuff
> ifdef'd out[1] and will try to build packages kdeedu3, kdebase4,
> games/wesnoth to see if there's any adverse side-effects.
>
> --
> Mel
>
> [1]
> --- Makefile.orig       2009-07-28 03:44:50.0 -0800
> +++ Makefile    2009-07-30 07:19:26.0 -0800
> @@ -25,9 +25,11 @@
>                PYTHON_INCLUDES="${PYTHON_INCLUDEDIR}"\
>                PYTHON_LIB_PATH="${PYTHON_LIBDIR}"
>
> -RUN_DEPENDS+=  
> ${PREFIX}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
> -RUN_DEPENDS+=  gccxml:${PORTSDIR}/devel/gccxml \
> -               
> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
> +RUN_DEPENDS+=  
> ${LOCALBASE}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
> +RUN_DEPENDS+=  
> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
> +.if defined(WITH_PYSTE)
> +RUN_DEPENDS+=  gccxml:${PORTSDIR}/devel/gccxml
> +.endif
>
>  BUILD_DEPENDS+=        ${BJAM}:${PORTSDIR}/devel/boost-jam
>  BUILD_DEPENDS+=        
> ${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
> @@ -39,12 +41,13 @@
>
>  post-patch: customize-boost-build
>
> -
> +.if defined(WITH_PYSTE)
>  post-configure:
>  # Configure pyste, the Boost.Python code generator
>       �...@cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
>        ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>                config ${PYDISTUTILS_CONFIGUREARGS}
> +.endif
>
>  do-build:
>  # build the library
> @@ -52,11 +55,12 @@
>                ${SETENV} ${MAKE_ENV} ${BJAM} ${BJAM_OPTIONS} 
> --prefix=${PREFIX}\
>                --toolset=${BOOST_TOOLS} --with-python
>
> +.if defined(WITH_PYSTE)
>  # Build pyste, the Boost.Python code generator
>        cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
>        ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>                build ${PYDISTUTILS_BUILDARGS}
> -
> +.endif
>  # Remove non-python headers
>        cd ${WRKSRC} && \
>        ${FIND} boost -not -path '*python*' -delete
> @@ -79,10 +83,12 @@
>        ${FIND} boost -type f -a \! -name "*.orig"\
>                -exec ${INSTALL_DATA} \{\} ${PREFIX}/include/\{\} \;
>
> +.if defined(WITH_PYSTE)
>  # Install pyste, the Boost.Python code generator
>        cd ${WRKSRC}/libs/python/pyste/install && \
>        ${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
>                install ${PYDISTUTILS_INSTALLARGS}
> +.endif
>
>  post-install:
>  # display pkg-message
>
>
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"


Re: devel/boost: new patch for testing update to 1.39

2009-07-30 Thread Mel Flynn
On Thursday 30 July 2009 03:20:56 Alexander Churanov wrote:

> I do not understand.
>
> Is it correct that you are discussing the issue of Pyste being part of
> boost and depending on gccxml port, which has issues of its own?
>
> I do not use Pyste and did not performed any specific testing of it,
> but all of boost-python-libs builds and installs, including
> dependencies.
>
> Please, describe what do you expect from ports and what do you
> observe, what negative outcome the issue has.

The previous boost-python port had an *optional* dependency on pyste.
The current one does not. This means that boost-python now pulls in
gccxml which pulls in lang/gcc34 on anything but 6.x.
I'm currently building boost-python-libs with the pyste stuff
ifdef'd out[1] and will try to build packages kdeedu3, kdebase4,
games/wesnoth to see if there's any adverse side-effects.

-- 
Mel

[1]
--- Makefile.orig   2009-07-28 03:44:50.0 -0800
+++ Makefile2009-07-30 07:19:26.0 -0800
@@ -25,9 +25,11 @@
PYTHON_INCLUDES="${PYTHON_INCLUDEDIR}"\
PYTHON_LIB_PATH="${PYTHON_LIBDIR}"
 
-RUN_DEPENDS+=  
${PREFIX}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
-RUN_DEPENDS+=  gccxml:${PORTSDIR}/devel/gccxml \
-   
${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
+RUN_DEPENDS+=  
${LOCALBASE}/lib/libboost_thread.so.${BOOST_SHARED_LIB_VER}:${PORTSDIR}/devel/boost-libs
+RUN_DEPENDS+=  
${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
+.if defined(WITH_PYSTE)
+RUN_DEPENDS+=  gccxml:${PORTSDIR}/devel/gccxml
+.endif
 
 BUILD_DEPENDS+=${BJAM}:${PORTSDIR}/devel/boost-jam
 BUILD_DEPENDS+=
${PYTHON_SITELIBDIR}/elementtree/ElementTree.py:${PORTSDIR}/devel/py-elementtree
@@ -39,12 +41,13 @@
 
 post-patch: customize-boost-build
 
-
+.if defined(WITH_PYSTE)
 post-configure:
 # Configure pyste, the Boost.Python code generator
@cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
config ${PYDISTUTILS_CONFIGUREARGS}
+.endif
 
 do-build:
 # build the library
@@ -52,11 +55,12 @@
${SETENV} ${MAKE_ENV} ${BJAM} ${BJAM_OPTIONS} 
--prefix=${PREFIX}\
--toolset=${BOOST_TOOLS} --with-python
 
+.if defined(WITH_PYSTE)
 # Build pyste, the Boost.Python code generator
cd ${BUILD_WRKSRC}/libs/python/pyste/install && \
${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
build ${PYDISTUTILS_BUILDARGS}
-
+.endif
 # Remove non-python headers
cd ${WRKSRC} && \
${FIND} boost -not -path '*python*' -delete
@@ -79,10 +83,12 @@
${FIND} boost -type f -a \! -name "*.orig"\
-exec ${INSTALL_DATA} \{\} ${PREFIX}/include/\{\} \;
 
+.if defined(WITH_PYSTE)
 # Install pyste, the Boost.Python code generator
cd ${WRKSRC}/libs/python/pyste/install && \
${SETENV} ${MAKE_ENV} ${PYTHON_CMD} ${PYSETUP} \
install ${PYDISTUTILS_INSTALLARGS}
+.endif
 
 post-install:
 # display pkg-message

___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"


Re: devel/boost: new patch for testing update to 1.39

2009-07-30 Thread Alexander Churanov
Mel,

I do not understand.

Is it correct that you are discussing the issue of Pyste being part of
boost and depending on gccxml port, which has issues of its own?

I do not use Pyste and did not performed any specific testing of it,
but all of boost-python-libs builds and installs, including
dependencies.

Please, describe what do you expect from ports and what do you
observe, what negative outcome the issue has.

Sincerely,
Alexander Churanov,
maintainer of devel/boost-*


2009/7/30 Mel Flynn :
> On Monday 29 June 2009 05:44:20 Alexander Churanov wrote:
>
>> I've made several changes to new set of boost ports. Please, review
>> them and test new set of ports, if possible.
>
> I guess I should've tested these, but I didn't expect it to replace a minor
> CONFLICTS problem that I had locally patched with a few lines, with a major
> not optional dependency on a retired compiler.
>
> Is pyste going to be optional in a new revision or is there any chance gccxml
> can be patched so it builds cleanly with the current GCC?
>
> If the Makefile comment is the only problem, surely it can be addressed?
>
> # GCC-XML has some constructs that GCC 4.2 does not like. Namely,
> # the use of casted-pointer-dereferences as an lvalue to post-increment
> # operations.
>
> Additionally the gccxml port seems pretty much dead. An attempt of revival has
> been done in 2007, but no release has ever been made.
> --
> Mel
>
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"


Re: devel/boost: new patch for testing update to 1.39

2009-07-29 Thread Mel Flynn
On Monday 29 June 2009 05:44:20 Alexander Churanov wrote:

> I've made several changes to new set of boost ports. Please, review
> them and test new set of ports, if possible.

I guess I should've tested these, but I didn't expect it to replace a minor 
CONFLICTS problem that I had locally patched with a few lines, with a major 
not optional dependency on a retired compiler.

Is pyste going to be optional in a new revision or is there any chance gccxml 
can be patched so it builds cleanly with the current GCC?

If the Makefile comment is the only problem, surely it can be addressed?

# GCC-XML has some constructs that GCC 4.2 does not like. Namely,
# the use of casted-pointer-dereferences as an lvalue to post-increment
# operations.

Additionally the gccxml port seems pretty much dead. An attempt of revival has 
been done in 2007, but no release has ever been made.
-- 
Mel
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"


Re: devel/boost: new patch for testing update to 1.39

2009-07-02 Thread Bruce Simpson

Thanks for this much needed work.
Detailed diagnostic info and patches follow.

System under test:
%%%
FreeBSD anglepoise.lon.incunabulum.net 7.2-STABLE FreeBSD 7.2-STABLE #6: 
Wed Jun 10 09:58:47 BST 2009 
r...@anglepoise.lon.incunabulum.net:/home/obj/usr/src/sys/ANGLEPOISE  amd64

%%%
Note: kernel+world ARE out of sync, world ahead by ~2 weeks, due to some 
catastrophic USB regressions in RELENG_7 affecting this machine.


I got a few warnings during 'make index'  which indicate there are 
stragglers:


%%%
anglepoise# make index
Generating INDEX-7 - please wait..make_index: avogadro-0.9.6: no entry 
for /usr/ports/devel/boost-python

make_index: avogadro-0.9.6: no entry for /usr/ports/devel/boost-python
^^^ boost-python in LIB_DEPENDS

make_index: peekabot-0.6.0: no entry for /usr/ports/devel/boost
make_index: peekabot-0.6.0: no entry for /usr/ports/devel/boost
^^^ boost in LIB_DEPENDS

make_index: flyhard-0.42: no entry for /usr/ports/devel/boost
^^^ boost in BUILD_DEPENDS

Warning: Duplicate INDEX entry: cvsup-without-gui-16.1h_4
Warning: Duplicate INDEX entry: mod_rpaf-ap2-0.6
^^^ these look unrelated
Done.
%%%

I've attached a diff for the stragglers, this fixes the 'make index' 
warnings.


%%%
anglepoise# portupgrade -o devel/boost-libs devel/boost
[Updating the portsdb  in /usr/ports ... - 20370 port 
entries found 
.1000.2000.3000.4000.5000.6000.7000.8000.9000.1.11000.12000.13000.14000.15000.16000.17000.18000.19000.2... 
. done]

cd: can't cd to /usr/ports/devel/boost
[chose VERBOSE_BUILD, DEBUG, ICU from options]
%%%

reinstall ok.  now upgrade dependent ports installed on my system:

%%%
anglepoise# portupgrade -rfx devel/boost devel/boost
%%%

no-op. try something else:

%%%
anglepoise# portupgrade -rfx devel/boost-libs devel/boost-libs
===>  Extracting for boost_build-2.0.m12
=> MD5 Checksum OK for boost-build-2.0-m12.tar.bz2.
=> SHA256 Checksum OK for boost-build-2.0-m12.tar.bz2.
===>  Patching for boost_build-2.0.m12
===>  Applying FreeBSD patches for boost_build-2.0.m12
===>  Configuring for boost_build-2.0.m12
--->  Backing up the old version
--->  Uninstalling the old version
--->  Deinstalling 'boost_build-2.0.m12'
pkg_delete: unable to completely remove directory 
'/usr/local/share/boost_build/tools'
pkg_delete: unable to completely remove directory 
'/usr/local/share/boost_build'

pkg_delete: couldn't entirely delete package (perhaps the packing list is
incorrectly specified?)
[Updating the pkgdb  in /var/db/pkg ... - 945 packages 
found (-1 +0) (...) done]

--->  Installing the new version via the port
===>  Installing for boost_build-2.0.m12
===>   boost_build-2.0.m12 depends on file: /usr/local/bin/bjam - not found
===>Verifying reinstall for /usr/local/bin/bjam in 
/usr/ports/devel/boost-jam

===>  Vulnerability check disabled, database not found
===>  Extracting for boost-jam-1.39.0
=> MD5 Checksum OK for boost_1_39_0.tar.bz2.
=> SHA256 Checksum OK for boost_1_39_0.tar.bz2.
===>  Patching for boost-jam-1.39.0
===>  Configuring for boost-jam-1.39.0
===>  Building for boost-jam-1.39.0
cd /usr/ports/devel/boost-jam/work/boost_1_39_0/tools/jam/src && 
./build.sh gcc;

###
### Using 'gcc' toolset.
###
rm -rf bootstrap
mkdir bootstrap
gcc -o bootstrap/jam0 command.c compile.c debug.c expand.c glob.c hash.c 
hdrmacro.c headers.c jam.c jambase.c jamgram.c lists.c make.c make1.c 
newstr.c option.c output.c parse.c pathunix.c pathvms.c regexp.c rules.c 
scan.c search.c subst.c timestamp.c variable.c modules.c strings.c 
filesys.c builtins.c pwd.c class.c native.c w32_getreg.c modules/set.c 
modules/path.c modules/regex.c modules/property-set.c modules/sequence.c 
modules/order.c execunix.c fileunix.c

./bootstrap/jam0 -f build.jam --toolset=gcc --toolset-root= clean
...found 1 target...
...updating 1 target...
...updated 1 target...
./bootstrap/jam0 -f build.jam --toolset=gcc --toolset-root=
...found 47 targets...
...updating 2 targets...
[MKDIR] bin.freebsdx86_64
[COMPILE] bin.freebsdx86_64/bjam
...updated 2 targets...
===>  Installing for boost-jam-1.39.0
===>   Generating temporary packing list
/bin/mkdir -p /usr/local/bin
install  -s -o root -g wheel -m 555 
/usr/ports/devel/boost-jam/work/boost_1_39_0/tools/jam/src/bin.freebsdx86_64/bjam 
/usr/local/bin/

===>   Registering installation for boost-jam-1.39.0
===>   Returning to build of boost_build-2.0.m12
===>   Generating temporary packing list

To use boost_build, it is necessary to set BOOST_ROOT in
your environment:

sh/bash/ksh:
   BOOST_ROOT=/usr/local/share/boost_build ; export BOOST_ROOT
csh:
   setenv BOOST_ROOT /usr/local/share/boost_build

Note: The default toolchain is gcc.
To override this on a system-wide basis, edit this file:
   /usr/local/share/boost_build/site-config.jam

For an example of a per-user settings file which may be

devel/boost: new patch for testing update to 1.39

2009-06-29 Thread Alexander Churanov
Hi folks!

I've made several changes to new set of boost ports. Please, review
them and test new set of ports, if possible.

The updated diff is at
http://alexanderchuranov.com/boost-port/boost-2009-06-29-14-38.diff
The updated tarball is at
http://alexanderchuranov.com/boost-port/boost-ports-200906291349.tar.bz2

The procedure is as follows:

1) Download the tarball and unpack under /usr/ports/devel
2) Go to /usr/ports, download and apply the patch
3) Remove 'devel/boost' and 'devel/boost-python' ports
4) Rebuild ports

The changes are:

1) Old ports removed from devel/Makefile, new added
2) URI for downloading source tarball is corrected
3) Python versioning issue in boost-python-libs is fixed
4) Minor changes in Makefiles of ports that were recently updated

Sincerely,
Alexander Churanov,
maintainer of devel/boost-*
___
freebsd-ports@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"