[Olena-patches] last-svn-commit-559-ge445824 Comment on an issue of Swilena highlighted by max-tree.py.

* python/max-tree.py: Here. --- swilena/ChangeLog | 6 ++++++ swilena/python/max-tree.py | 14 ++++++++++++-- 2 files changed, 18 insertions(+), 2 deletions(-) diff --git a/swilena/ChangeLog b/swilena/ChangeLog index 868a17e..38e7be9 100644 --- a/swilena/ChangeLog +++ b/swilena/ChangeLog @@ -1,5 +1,11 @@ 2012-08-28 Roland Levillain &lt;roland(a)lrde.epita.fr&gt; + Comment on an issue of Swilena highlighted by max-tree.py. + + * python/max-tree.py: Here. + +2012-08-28 Roland Levillain &lt;roland(a)lrde.epita.fr&gt; + Update the local (build-time) Swilena Python Shell. * python/sps-local.in (PYTHONPATH): Add `@abs_builddir@/.libs'. diff --git a/swilena/python/max-tree.py b/swilena/python/max-tree.py index 05dd32a..236c75d 100644 --- a/swilena/python/max-tree.py +++ b/swilena/python/max-tree.py @@ -1,6 +1,6 @@ #! /usr/bin/env python -# Copyright (C) 2010 EPITA Research and Development Laboratory (LRDE) +# Copyright (C) 2010, 2012 EPITA Research and Development Laboratory (LRDE) # # This file is part of Olena. # @@ -37,8 +37,18 @@ for p, v in itertools.izip(ima.domain(), values): ima.set(p, int_u8(v)) image.println("ima =", ima) +# FIXME: The type of the (Python) object returned by this call, +# <type 'SwigPyObject'>, is wrong. The right type should be +# <class 'image2d_point2d.image2d_point2d'>, but the fact that the +# routine and the result object are located in two different +# (generated) Swilena submodules seems to confuse SWIG. A consequence +# of this issue is that `max_tree_parent' is not usable at all, hence +# the failure of the last line of this test. We should perhaps review +# Swilena's design and minimize or even completely get rid of the +# modular approach... max_tree_parent = image.max_tree(ima, c4()); -# FIXME: Why can't we use + +# FIXME: Overloading issue: Why can't we use # # image2d_point2d.println("max_tree_parent =", max_tree_parent) # -- 1.7.2.5