On Mon, Aug 20, 2001 at 02:18:17AM -0700, Guy Harris wrote: > I also checked in a change to free up the exception ID string if one was > allocated; it needs some additional work - you need to catch exceptions, > so that if a dissector throws an exception, we still free the string, > and then the exception has to be rethrown. > > Either that or we need to add a cleanup callback to free the string - > and any *other* dynamically-allocated data.... I've checked in a change that moves a bunch of "g_free()" calls as close as possible to the allocation, which, if there are no tvbuff references between the allocation and the free, obviates the need for a cleanup callback. (I found the calls by auditing the code, which also found some places where there *were* no "g_free()" calls, and one place where the allocations were happening in a loop but the free was happening outside the loop.) In the remaining cases, where there had to be tvbuff references between the allocation and the freeing, I added cleanup callbacks.
Powered by MHonArc 2.6.10