On 06/01/17 13:35 +0000, Jonathan Wakely wrote:
I'm surprised we don't have any tests for this case, but apparently we don't, as your patch passes all our tests.
My bad, we do have tests that FAIL with this patch, but only after a 'make clean' (otherwise the explicit instantiation definitions in the library don't get rebuilt after applying the patch, so the tests use the old code). FAIL: 21_strings/basic_string/operations/find/char/1.cc execution test FAIL: 21_strings/basic_string/operations/find/wchar_t/1.cc execution test These fail on assert(s.find(s, 1) == std::string::npos) which gives the wrong answer as explained in my previous mail. The infinite loop caused by this patch isn't caught by any existing tests, so I'll add a check for that.
