On Mon, Sep 09, 2013 at 05:08:31PM +0200, Simon Wunderlich wrote: > Hrm, apart from that this patch lacks a commit message and signoff, what does > it help > freeing stuff just before the program exits? This will free anything > automatically > anyway ... >
Hm, the advantage of freeing such memory in the code is that for one thing it makes valgrind happy :P. And this in turn makes finding "real" memory leaks easier. And besides, isn't it a better coding style to free such global variables, too? I think I'd love seeing a PATCHv2 of this one. And I guess only a short, two sentence commit message should be sufficient as this isn't a very invasive patch. For instance: "This patch fixes a non-serious memory leak on program exit by freeing the memory allocated for the globals variables. This makes valgrind happy, too. Signed-off-by: ..." Cheers, Linus PS: Sorry, I had a glance over this patch before it got submitted but even though I should be somehow familiar with the prefered coding style, I didn't object.
