[patch] (Lenka's) Improved handling of unsuccessful adaptivity steps in adapt()

[GoogleCodeExporter]

After introducing p-adaptivity, there was a new problem that the elements
selected for refinement could not be refined because the maximum order was
reached. This situation was not handled correctly, Hermes did not refine
any element but said that he refined some elements. Now there is a warning,
Hermes says that he refined 0 elements, and the adapt() function returns
'true' that the user can handle and stop the computation if he likes. 

Original issue reported on code.google.com by solin.pa...@gmail.com on 10 Aug 2009 at 4:14

Attachments:

• 0001-Handling-of-unsuccessful-adaptivity-steps-improved.patch

[GoogleCodeExporter]

This patch does not apply on the master branch. Check the file 
src/adapt_ortho_h1.cpp 
from the master branch in the upstream for "successfully_refined" and you won't 
find 
any hit. So it seems that some of the modifications you did were not merged 
into this 
patch. Please fix that and resend it. Thanks.

I'm removing NeedsReview tag. Once the new patch is submitted, please, set it 
again.

Original comment by and...@gmail.com on 23 Aug 2009 at 1:48

• Removed labels: NeedsReview

[GoogleCodeExporter]

Closing this as it was solved off-list off-issue tracker.

Original comment by and...@gmail.com on 23 Aug 2009 at 2:32

• Changed state: Fixed

[GoogleCodeExporter]

Please solve things either on list, or on the issue tracker.

Original comment by ondrej.c...@gmail.com on 23 Aug 2009 at 2:48