[vlc-devel] Re: vlc: svn commit r18218 (courmisch)
Sigmund Augdal Helberg
dnumgis at videolan.org
Sun Dec 3 21:45:43 CET 2006
On 12/3/06, Clément Stenac <zorglub at videolan.org> wrote:
> >> Also can you say a few words on when/how this memleak happened and
> >> an approximation on how log it's been around?
> It happened last week when I made most of input_thread_t private.
> input_thread_t is now basically
> a few public things;
> input_thread_private_t* p;
> ... and I forgot to free it ...
> So the memleak was a few hundreds of bytes at each input thread, since
> last week.
Ok, nice to know.
>> And finally, shouldn't p_input->p get a proper name?
> > Probably (but *I* won't do it).
> I actually didn't want a p_very_long_name as it appears extremeley
> often, but it wouldn't be hard to change that.
I that there is likely a relation between the variables short name and
it not being freed.
People would tend to think "this 'p'-variable can't possibly be important,
so I will not check it up." I'd say anything this important deserve a proper
name, and p_private isn't that long. Functions using it a lot could maintain
a shorthand anyway like most things do for foo->p_sys. Having core code use
so short names will make the source base harder to maintain and harder to
understand for new coders.
> This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
> To unsubscribe, please read http://developers.videolan.org/lists.html
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the vlc-devel