Comment 3 for bug 1251687

Rebasing on top of trunk, and doing some trivial whitespace cleanup to make the diff a little more readable....

Shouldn't we be calling g_queue_free_full rather than g_queue_free to avoid leaking the g_new0'd trackpoint_t objects? There's not a race or anything that we're avoiding by not freeing them, is there?

Also, I don't quite understand what you're doing with the NULL sentinels in tracks.c and callbacks.c--can you provide some comments?