[ET Trac] [Einstein Toolkit] #1186: Provide a version of CCTK_WARN which never returns

[Einstein Toolkit]

#1186: Provide a version of CCTK_WARN which never returns
--------------------------+-------------------------------------------------
  Reporter:  hinder       |       Owner:     
      Type:  enhancement  |      Status:  new
  Priority:  minor        |   Milestone:     
 Component:  Cactus       |     Version:     
Resolution:               |    Keywords:     
--------------------------+-------------------------------------------------

Comment (by knarf):

 This sounds like a good idea. However, is it known whether other compilers
 than gcc honor this attribute? If not this would only silence the warning
 if gcc is used. However, there could be a new macro like the proposed
 CCTK_ERROR which is actually calling 'exit()' after CCTK_WARN(0) (which
 would never be really called, but the compiler would see it). If we go
 that way, CCTK_VError would also me handy, analogous to CCTK_VWarn.

 If one would like to be picky then one could also point out that a
 negative error level is allowed in Cactus (but probably shouldn't, which
 is another issue). That means CCTK_WARN(0,"") does in fact not necessarily
 abort right now (but again, probably should).

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/1186#comment:3>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit