On 8-6-2012 16:29, Jon Cruz wrote:
Hi,
I finished a simple writeup of converting C strings to C++ strings for our codebase. A couple of people asked me for this, but I thought it might be more generally useful.
A shorter summary probably will be appropriate to get on the wiki at some point, but for now
http://codewideopen.blogspot.com/2012/06/refactoring-c-to-c-part-2-strings.h...
Hi Jon, Nice that you took the time to write that up. I also read part1 of your C to C++ series [1] and have a comment and a question.
The article reads: "Another handy aspect to turning stand-alone C functions in to C++ methods is that we get compile-type checks and safety and can drop run-time checks, such as at the beginning of the new SPCtrlLine::setRgba32() method: 154 g_return_if_fail (cl != NULL); 155 g_return_if_fail (SP_IS_CTRLLINE (cl));" and recommends we drop these two lines. There is a subtle change going from sp_classname_set(class*) to class->set(). From a callers perspective, the former seems to accept NULL, whereas the latter does not. If the caller relied on NULL pointer handling, at least the first line should be kept, or the calling code should be adjusted. There have been a number of crashers that involved class methods being called with this=NULL. It is easy to recognize such crashes, and a fix is easy. I agree we can remove those lines, but not when release is getting close...!
Question: For me a big benefit of C++ classes is virtual methods (notably virtual destructors!). Any way we can get rid of the very ugly and bugprone parent_class stuff?
Cheers, Johan
[1] http://codewideopen.blogspot.ch/2012/05/refactoring-c-to-c-part-1.html