[Users] [svn:cactuscode] Boundary/trunk/src/ (Rev. 327)

[Roland Haas]

Hello Erik, all,

> It seems that these lines pertain to dimensions that are larger than
> the dimension of the respective grid variable. That is, this would
> e.g. set the extent of the z (sic!) boundary to 1 for 2d variables
> that don't have a z direction. Is this intended?
Yes, it treats a 2d grid variable as a 3d grid variable with 1 point in
the z direction. This should be fine as far as I know.

Before revision 309 (which introduced ash and removed lssh) the code
actually read:
--8<--
   for (i = gdim; i < MAXDIM; i++)
   {
    lssh[i] = 1;
    lsh[i]  = 0;
+    lsh[i] = 0;
   }
--8<--
which 309 changed to
--8<--
   for (i = gdim; i < MAXDIM; i++)
   {
    lsh[i] = 0;
   }
--8<--
which is actually wrong since it will mean eg the kk loop in
SCALAR_BOUNDARY_TYPED (in the scalar boundary code) will never execute
its loop body and the b/c will not be applied anywhere.

Instead of setting ash to 1 one could also set it to 0 but that leads to
the situation that ash < lsh which is no the case for regular grid
variables.

Yours,
Roland

-- 
My email is as private as my paper mail. I therefore support encrypting
and signing email messages. Get my PGP key from http://keys.gnupg.net.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 242 bytes
Desc: OpenPGP digital signature
Url : http://lists.einsteintoolkit.org/pipermail/users/attachments/20140429/f9deeae2/attachment.bin