Comment # 16
on bug 6020
from Hadriel Kaplan
One update on my comment - the first pass vs. the second pass was reversed. I
had put in a print statement in the Lua script which made it appear this way,
but digging into the code showed it otherwise.
The specific fields being extracted (the ones from Fields.new()) apparently
*are* filled-in in this first pass, but only those fields - because the
underlying C-code creates a fake tap with a fake dfilter for those fields,
presumably to make this exact scenario possible.
Unfortunately, the fake tap isn't setting the TL_REQUIRES_COLUMNS flag, so
during the first pass columns are not being made accessible. For whatever
reason, however, during the second pass (when columns are accessible), the
particular tree fields are not being populated - some are, most are not until
the specific packet is selected, but by then columns are inaccessible again.
But this behavior is inconsistent packet-by-packet, which is why their example
scripts succeeded to change the protocol column text sometimes but not others.
Weird.
You are receiving this mail because:
- You are the assignee for the bug.
- You are watching all bug changes.
