yade-dev team mailing list archive

[Václav Šmilauer <eudoxos@xxxxxxxx>]

Nice plans, some points, though.

For the regression tests, there is no single reference test in yade
proper, so how are you going to test that? (besides that, I don't see
how serialization could affect test results, except that the XML file is
badly written/read, which can be checked IMHO rather easily by hand).

If boost::extension is needed, we can have it in-tree and then just
switch to the packaged version, once it is available in boost. The same
BTW for loki, which we don't really need inside yade anymore, I think?

> 5. Perhaps implement a "serialization on demand" interface. It uses
>    boost::serialization, and can be used to check values when their name
>    is given. The name is provided using syntax similar to directory/file.
>    So you would ask gimmie("SomeClass/SomeMember/torsionalStiffness");
>    and get the answer - the value. That could be later used for tracking
>    variables maybe obtaining simulation results, and debugging.
>   
Good, python wrapper currently modifies all attributes by
serializing/deserializing them and until c++ classes are reflexted in
python directly (using py++, perhaps), I would like to keep this
functionality.

For the qt4 stuff: I think that many people would be happier using
scripting interface than clickable interface (I prefer it very much: you
can reproduce what you do exactly, save it to file etc). WRT qt3 gui,
the missing part of python now is 3d-view. It freezes X on my notebook,
for unknown reasons. If you dare, go try it on yours
(gui/cmd/yadeControl.cpp, class pyView), perhaps it is some threading
issue that I don't understand. I would be gald if you could help me with
that - I don't like hard-rebooting my computer. Perhaps it is only a
driver issue (realtively new i965), it used to work on ATI.

Vaclav
_______________________________________________
yade-dev mailing list
yade-dev@xxxxxxxxxxxxxxxx
https://lists.berlios.de/mailman/listinfo/yade-dev