Carsten Burstedde
Carsten Burstedde
If malloc returns NULL, we are out of memory. This seems strange with a relatively small mesh. You might call some of the *_memory_used functions to investigate the memory consumption...
Thanks for checking and verifying that all other tetst are ok. So the allocation is for 2.8G, understandable it crashes on 32bit.
Thanks, those are all good points. If you'd like to make a PR against the develop branch, that would be great. Exposing `p4est_vtk_write_cell_datav` sounds good as well as implementing `p4est_vtk_write_cell_data`....
Thanks for looking into it! > I went with copy of `p4est_vtk_write_cell_dataf` where I changed variadic arguments for va_list. I thought about exposing the `p4est_vtk_write_cell_datav`, however there are some checks...
Looks great then! We're still missing the point analogons for data and datav. Contributions welcome.
Additionally: * Add a new CI job that runs cpack, unpacks the archive und then runs the autotools chain. * Add a new CI job that calls make dist, unpacks...
Another point: * The tests on make check run with 2 MPI ranks. Make ctest behave the same if not so.
A couple more items to align the CMake to the automake build: * Currently, make install without explicit prefix installs into sc/local. Change to local. * When installing and sc...
* `make install` creates uppercase SC and P4EST files/directories. Preferred convention would be lowercase.
> I think to make "examples" be a full test of the "installed" p4est from a single project instead of two separate projects as I have now would require making...