Hi Jiri, i'd be happy to accept such patch.
Best regards, Dimitri
On Sun, 04 Oct 2015 15:58:53 +0200, P j-pi@seznam.cz wrote:
Hi,
there is in gr-osmosdr a lot of code like:
hackrf_source_c::~hackrf_source_c () { ... free(_buf); _buf = NULL; } }
Setting value in destructor have sense in two case:
- they are used in some way in destructor of parent class
- You set then to something like 0xDEADBEEF and when pointer from already destroyed class is used it will cause segfault.
Setting value to NULL is contraproductive because there is a lot of checks if value is not NULL, this actually can hide some troubles of calling something using already freed pointer.
I'm suggesting to remove this kind of code. I will prepare patch but now I'm waiting if those two pending patches I have send before would be accepter or not.
Is there any opinion which oppose previous claim?
With best regards, JP.