Hi, if the memory leaked from _mesa_malloc during a glTexSubImage2D call, it seems more likely to me that it is a Mesa bug during the discarding of the old texture and the compositing of the new one.
Just my thoughts ;) Alberto El Viernes, 2 de Febrero de 2007 10:58, Panagiotis Papadakos escribió: > There seems to be a memory leak in osgText. The output of valgrind it the > following: > > ==4911== 2,700 bytes in 675 blocks are definitely lost in loss record 214 > of 319 > ==4911== at 0x40234E8: malloc (vg_replace_malloc.c:207) > ==4911== by 0x4B6C874: _mesa_malloc (imports.c:79) > ==4911== by 0x4B8F130: _mesa_init_teximage_fields (teximage.c:1183) > ==4911== by 0x4B73830: _mesa_generate_mipmap (mipmap.c:936) > ==4911== by 0x4B9C2CE: _mesa_store_texsubimage2d (texstore.c:3181) > ==4911== by 0x4B304E6: r300TexSubImage2D (r300_tex.c:768) > ==4911== by 0x4B91B18: _mesa_TexSubImage2D (teximage.c:2696) > ==4911== by 0x47CB8E7: glTexSubImage2D (glapitemp.h:1828) > ==4911== by 0x44E5221: osgText::Font::Glyph::subload() const > (Font.cpp:718) ==4911== by 0x44E8EE4: > osgText::Font::GlyphTexture::apply(osg::State&) const (Font.cpp:570) > ==4911== by 0x44FEBD2: osg::State::applyTextureAttribute(unsigned, > osg::StateAttribute const*) (State:927) > ==4911== by 0x44F6595: > osgText::Text::renderWithPolygonOffset(osg::State&, osg::Vec4f const&) > const (Text.cpp:1986) _______________________________________________ osg-users mailing list [email protected] http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
