* origin/topic/robin/pp-alarms:
The silliest, tiniest little whitespace fixes.
Update missing in last commit to this branch.
Adding test for alarm mail.
Tuning the pretty-printed alarms output.
* origin/topic/seth/notice-email-delay:
The hostname notice email extension works now.
Fixed more bugs with delayed emails.
Working around a problem with setting default container types.
Ugh, still major failure. I'm just cutting the timeout handling for now.
Fixed a small bug major problem with email delay timeout catching.
Initial fixes for the problem of async actions with notice email extensions.
Closes#727.
* origin/fastpath:
Fix missing action in notice policy for looking up GeoIP data.
Better persistent state config warning messages (fixes#433).
A few updates for SQL injection detection.
Fixed some DPD signatures for IRC. Fixes ticket #311.
Removing Off_Port_Protocol_Found notice.
SSH::Interesting_Hostname_Login cleanup. Fixes#664.
Teach Broxygen to more generally reference attribute values by name.
Fixed a really dumb bug that was causing the malware hash registry script to break.
Fix Broxygen confusing scoped id at start of line as function parameter.
Remove remnant of libmagic optionality
- The biggest change is the change in notice names from
HTTP::SQL_Injection_Attack_Against to
HTTP::SQL_Injection_Victim
- A few new SQL injection attacks in the tests that we need to
support at some point.
- Added a field named $last_alert to the SSL log. This doesn't even
indicate the direction the alert was sent, but we need to start somewhere.
- The x509_certificate function has an is_orig field now instead of
is_server and it's position in the argument list has moved.
- A bit of reorganization and cleanup in the core analyzer.
The loading of these is better handled by BroControl and it seems
odd to load them from a base/ script anyway since they'll contain
site/policy specific code.
Addresses #663
- Answers and TTLs are now vectors.
- The warning that was being generated (dns_reply_seen_after_done)
from transaction ID reuse is fixed.
- Updated the single failing btest baseline.
The only snag is... with the default output format of the log-file writer, the input reader cannot tell if a table or set is empty or uninitialized (both cases use the same character by default). In this case, by default it is assumed that the field/vector is uninitalized.
