Comment # 2
on bug 11863
from Michael Mann
The CIP dissector calls tvb_new_real_data to create a tvb based on some saved
data from the request (stored as a byte array), to show in the response packet.
This is effectively "generated" data, but since the tvb exists,
proto_tree_add_item was used with offsets into the tvb created by
tvb_new_real_data. However, add_new_data_source() is never called, which is
why the PDML export asserts. It's not really the intention of the CIP
dissector to have a "new data source", it just wanted a tvb to call
proto_tree_add_xxx on for its generated items. Should the behavior then be (as
it is in the Gerrit patch) that the tvb from tvb_new_real_data can be used, but
you can't use proto_tree_add_item, you must use proto_tree_add_uint (or
similar) and always have an offset and length of 0?
You are receiving this mail because:
- You are watching all bug changes.