[Libmesh-devel] Policy on large files in CVS...

[John P. <pet...@cf...>]

I don't know what the filesize limits are (I seem to remember
something like 50MB) but I'd really rather not check in a huge file
like that to CVS.  We could host it from the CFDLab, and have a script
which wget's it when the example is run.

-J


Derek Gaston writes:
 > So... I'm modifying ex14 to include an option to compare against a
 > fine grid solution.  The problem is that the fine grid solution is
 > pretty large.  There are three files to go with it:
 > 
 > lshaped_fine_mesh.xda:  19MB
 > lshaped_fine_es.xda: 4.6MB
 > lshaped_fine.gmv: 13MB
 > 
 > The gmv isn't strictly necessary, but it is nice to see just how fine
 > of a solution you are comparing against.  This solution is a 7 times
 > uniform refinement of the original problem.
 > 
 > Do we want to check something like this in?  Should I just do less
 > uniform refinements?  The problem is that with this particular problem
 > we are going to probably get a better solution with adaptivity than
 > the fine grid pretty quickly (since most of the error is localized in
 > the singularity).
 > 
 > Any suggestions?
 > 
 > Derek