* mln/io/magick/load.hh,
	* mln/io/magick/save.hh: Use pointers to access image data.

	* mln/convert/to_qimage.hh: Fix a duplicate declaration.

	* mln/convert/to_qimage_nocopy.hh: Fix the minimum version
	required.

mln/convert/from_to.hh: Use value::cast for default builtin conversions and use specific from_to_ overloads for float/double conversions

	* tests/unit_test/disabled_tests: Add a missing new line at end of
	file.

	* tests/unit_test/unit-tests.mk: Regen.

	* tests/accu/stat/Makefile.am: Add a new target.

	* tests/accu/stat/deviation.cc: New.

	* mln/convert/from_to.hxx,
	* mln/convert/impl/from_unsigned_to_value.hh: Remove overload
	converting unsigned to float.

	* headers.mk,
	* tests/unit_test/unit-tests.mk: Regen.

	* mln/world/rgb/invert.hh,
	* mln/world/rgb/invert.spe.hh: New.

	* mln/arith/revert.hh,
	* mln/arith/revert.spe.hh: Add a standard dispatch.

	Yann's sandbox.

	* mln/convert/from_to.hxx,
	* mln/value/int_u.hh: Here.

	* mln/canvas/browsing/backdiagonal2d.hh
	* mln/canvas/browsing/diagonal2d.hh
	* mln/canvas/browsing/directional.hh
	* mln/canvas/browsing/snake_fwd.hh
	* mln/canvas/browsing/snake_generic.hh
	* mln/value/qt/rgb32.hh

Yann Jacquelet authored Jun 21, 2010 and Guillaume Lazzara committed Nov 15, 2010

	* mln/convert/from_to.hxx: New conversion utilities.
	* mln/convert/impl/from_unsigned_to_value.hh: New conversion utilities.

	* mln/io/magick/load.hh
	(mln::io::magick::load(Image<I>&, const std::string&))
	* mln/io/magick/save.hh
	(mln::io::magick::impl::get_color(const value::int_u8&))
	(mln::io::magick::impl::get_color(const value::rgb8&)):
	Properly convert values between Magick::Quantum and
	mln::value::int_u8, and vice versa.

	* tests/accu/Makefile.am (tuple_CPPFLAGS): New.
	* tests/unit_test/Makefile.am (BOOST_TUPLE_CPPFLAGS): New.

	* Makefile.am (dist_img_DATA): Add img/picasso.ppm.

	* tests/labeling/Makefile.am (check_PROGRAMS): Remove colorize.
	(colorize_SOURCES): Disable.

	* tests/value/builtin/floatings.cc: New.
	* tests/value/builtin/Makefile.am (check_PROGRAMS): Add floatings.
	(floatings_SOURCES): New.

	* mln/value/ops.hh (mln_trait_value_sum_product_)
	(mln_sum_product_): New macros.

	* mln/algebra/vec.hh (mln::algebra::vec<n,T>::normalize): Use
	mln_sum_product(T,T) instead of the hard-coded type `float' as
	return value of mln::norm::l2.

	* mln/fun/x2x/rotation.hh
	(fun::x2x::internal::get_rot_h_mat(const C, const algebra::vec<n,C>&)):
	Use std::abort instead of a precondition to stop the execution,
	since preconditions can be skipped by defining NDEBUG.

	* tests/fun/x2x/composed.cc: Fix the initialization of the axis
	vector.

	* mln/fun/x2x/rotation.hh
	(fun::x2x::internal::get_rot_h_mat(const C, const algebra::vec<3,C>&)):
	Fix the definition of the (2,1) coefficient.

	* tests/accu/Makefile.am: Remove uses of the old
	HAVE_BOOST_PREPROCESSOR conditional.
	* tests/unit_test/cond_tests_boost-preprocessor_boost-tuple:
	Rename as...
	* milena/tests/unit_test/cond_tests_boost-tuple: ...this.
	* tests/unit_test/Makefile.am (COND_TESTS):
	s/cond_tests_boost-preprocessor_boost-tuple/cond_tests_boost-tuple/.
	* tests/unit_test/unit-tests.mk: Regen.

	* mln/accu/label_used.hh,
	* mln/core/internal/labeled_image_base.hh,
	* mln/labeling/colorize.hh,
	* mln/labeling/compute.hh,
	* mln/labeling/superpose.hh,
	* mln/make/relabelfun.hh,
	* mln/set/compute_with_weights.hh: Make use of value::next.

	* mln/value/label.hh: Include next routine.

	* mln/value/next.hh: New.

	* mln/io/magick/load.hh,
	* mln/io/magick/save.hh: Here.