Bruno Haible wrote: > Hi Jim, > >> Bruno, is it ok to remove the one in lib/clean-temp.c, below? > > No, it's better if the 'if (old_array != NULL)' stays there. It reminds > the maintainer of the code of the two possible cases.
I'm surprised you think it's worth keeping that test solely as a reminder. > So, can you please add clean-temp.c to if_before_free_offenders_ ? I will. > Note also that this "avoid if before free" rule is less valuable nowadays > than a couple of years ago: It was a migration aid from K&R C to ANSI C. > Nowadays little code remains that is written for K&R C, and few minds program > for K&R C. Therefore the "if before free" cases that are remaining are > more likely to be intentional than before. If only that were true. I've just removed several hundreds of such uses from glusterfs. From what I've seen, the majority of C/C++ developers attempt to avoid free(p) when p is NULL.
