Tobias Burnus wrote:
The problem is fixed by the attached patch. I will commit it as obvious (to the trunk, i.e. GCC 5 alias GCC 4.10 only) once building and regtesting has finished.
As I only saw later in Steve's email, it is a regression. Thus, I will also apply it to the GCC 4.9 branch (will be in GCC 4.9.2; Linux distros likely pick it up earlier).
The regression was caused by my commit r197159 for PRs 56650 and 36437, which added compile-time simplification for storage_size, c_sizeof and sizeof. (The latter are not affected as they are supposed to return a value of kind c_size_t and take no kind parameters.)
Tobias
