Comment # 3
on bug 9047
from Roland Knall
(In reply to comment #2)
> Still need to do:
> 1. Make Payload values filterable. All of the "integer" types can share the
> same display filter name, but the "bytes" type will need to be different.
> Much better than a FT_NONE type.
> 2. If cleaning up, perhaps there are more unnecessary NULL checks for trees
> and items in the dissector. I only cleaned up ones in the provided patch.
The checks where part of the original dissector code, and I just added it to
the new parts. But I agree, they can stay out.
But can you precise the filtering of the bytes payloads? Do you want to have
different tags for WriteIndex, ReadIndex and WriteMultipleIndex? I disagree on
the base of the protocol, because the payload always contains the exact same
objects. Also, filtering by payload does not happen in reality. Filtering
happens by index/subindex, node id, content type, message type and so on.
Therefore I would keep the same sdo field index for all sdo payloads and simply
allow for a filter which basically says, that there exists a payload.
I will upload the new version tomorrow.
You are receiving this mail because:
- You are watching all bug changes.