REQUIRE assumes a char pointer is a valid string about catch2

[philsquared]

Thanks for reporting this, jafld.
Looks like the problem was just with the return of foo() - at least for me the literal NULL didn't crash (but it wasn't reported as nicely as it could be).

Anyway, I've fixed the issue (only tested with XCode GCC so far) - I just needed a couple of overloads of toString() that take char* const and const char* const, respectively.
Then I can handle the special case of NULL quite easily (which I now do, and print a special string: {null string}, in that case - and quote all other strings). This seems more friendly.
Let me know what you think and if it works correctly for you.

[jalfd]

Seems to work fine. Although I did bump into a related issue. If I try to compare against the C++0x nullptr instead of NULL, I get a compiler error (operator << is ambiguous).

Of course, you never made any claim of supporting C++0x, but thought I'd pass it on. It's no big deal, of course, since everything works fine if I use NULL.