Jeff Morriss <jeff.morriss.ws@...> writes:
>
>
> I hope so, but it appears to have worked before my change and not after,
> so... I'm a bit nervous.
>
Jeff:
Your intuition is correct. The current version of dumpcap is crashing on Windows
because of a null pointer in capture_opts->iface_descr and thus was causing the
Windows buildbot 'run tests' step to fail.
Unfortunately, the 'run tests' failure was somehow screwing up the buildbot
so it wasn't possible to look at the log for that step to see what was happening.
The specific problem: dumpcap calls capture_loop_start (in capture_loop.c) which
calls capture_opts_log (in capture_opts.c) which then passes a null pointer to
glog (for iface_descr); glog then causes dumpcap to die ....
A very quick look suggested I'd need to take a deeper look to determine a fix so
I decided I better leave this to you...
Bill