On 13 April 2015 at 03:13, Jed Brown <[email protected]> wrote: > Barry Smith <[email protected]> writes: > >> Lisandro, >> >> More code analysis is always good. We'll figure something out for this >> particular beast. Note that petscsys.h already uses #if >> defined(__has_attribute) so maybe we can use >> >> #if defined(__has_attribute) >> # if __has_attribute(analyzer_noreturn) >> >> instead of the #ifdef __clang_analyzer__? > > Should we add a PETSC_HAS_ATTRIBUTE(analyzer_noreturn) to make this less > verbose? (It would return 0 if __has_attribute is not available.) I > foresee more use of __has_attribute to reduce the number of tests we > have to run and make petscconf.h more robust to, e.g., minor version > changes in the compiler stack.
Yes, of course, this is much better. I guess petscsys.h is the right place to add this logic. I'll put this together in a PR. -- Lisandro Dalcin ============ Research Scientist Computer, Electrical and Mathematical Sciences & Engineering (CEMSE) Numerical Porous Media Center (NumPor) King Abdullah University of Science and Technology (KAUST) http://numpor.kaust.edu.sa/ 4700 King Abdullah University of Science and Technology al-Khawarizmi Bldg (Bldg 1), Office # 4332 Thuwal 23955-6900, Kingdom of Saudi Arabia http://www.kaust.edu.sa Office Phone: +966 12 808-0459
