Comment # 6
on bug 8343
from Evan Huus
(In reply to comment #5)
> Checked in possible fix in r47705. I say "possible" because I'm not seeing
> the crash on Windows to completely verify the fix.
The crash and valgrind errors are gone with r47705, so you were right on the
money.
> The issue is that there are 2 response packets for a single offer, and once
> an offer has been "processed" (accepted or rejected), the hash tables are
> freed. Subsequent response packets should be ignored (which is what r47705
> does).
>
> While I agree the safety net of SDP should be there regardless, it appears
> to me like there is a bug in the SIP dissector and that packet 208 should be
> marked as a resend of packet 194. If packet 208 was correctly marked as a
> "resend", the crash wouldn't have been there for SDP.
>
> Leaving bug open for explanation of SIP behavior, reducing bug severity.
Is it worth adding Anders or someone to the CC who might be able to explain?
You are receiving this mail because:
- You are watching all bug changes.
