my quick and dirty solution (see below) based on rapidxml works now
reliably, but you are right: It is an unnecessary effort parsing such
a simple xml file. It would be enough to write the whole file every
time anew (but then you have to store the timesteps somewhere else).
I will have a look on "write_pvtu_record" in "data_out_base.cc" in
the next weeks.
Great. Let us know if you have something!
Concerning the meta-data: with "part= " it should be possible to
specify different parts of a geometry or mesh for each timestep and
with "group=" you can arrange geometries in different groups - but I
am no expert in all this. It is just what I found on the net. As
described here (http://www.elmerfem.org/?p=223) the listing within
the ".pvd" file should work with ".pvtu" files for every time-step as
well (a kind of wrapper-wrapper).
OK. That would probably the simplest. In that case I'd say don't bother
with the part and group tags and make the function take an argument of
the kind
std::vector<std::pair<double,std::string> >
that indicates the pairs of current time and vtu/pvtu file information.
In user programs one would then just add one such pair to the end of the
vector in each time step. This is analogous to what one does for the
write_pvtu_record function.
Best
W.
------------------------------------------------------------------------
Wolfgang Bangerth email: [email protected]
www: http://www.math.tamu.edu/~bangerth/
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
