Documentation set which writes output in dia format.
Some known bugs/deficiencies
you can NOT specify an output filename yet, it is fixed to
dia.dia (quite easy to solve, I just haven't had the time). The
dia file is not yet compressed, but it can be used without changes
in dia. If you save the file from within dia, it will be saved in
gzipped format (without .gz). in some cases, the generalization arrows are not actually
connected to the class, but sometimes they will snap to the
right class when you move the class box on the screen. This
appears only (I think) if the name of the class in the dia class
font (which is a proportional font) is longer than the longest
method-name (which is displayed in a non proportional font). At
the moment I compute the size of the boxes only by calculating the
width of the string for the longest method name. I have no idea
how to get the length of the class name string :( ( I can work
around this one, if I connected the arrows to the left side of the
boxes instead of the middle, but I like it better that way.) If a base class is not included in the directories the dia file
is produced from, a message will be issues base_class <...> not
found and a dummy UML object is created (with just one __init__
method). The Layout is of course not really good (esp. for automatically
generated base classes, since they will be put in the same line
as the class where it is needed) No package support right now. Just a matter of time ;-) I assume that each class name is unique. If you have several
classes with the same name, one of the classes will have all the
generalizations arrows!
To Do
update an existing dia file if you change the python code
(e.g. just adding the new methods). automatically generate python code for an UML diagram? add attributes as well (matter of the parser)
|