Comment # 10
on bug 8581
from Evan Huus
(In reply to comment #9)
> - I don't see how 'attributes_follow' could possibly be uninitialized given
> the
> following code:
> attributes_follow = tvb_get_ntohl(tvb, offset+0);
> proto_tree_add_text(post_op_attr_tree, tvb, offset, 4,
> "attributes_follow: %s (%u)",
> val_to_str_const(attributes_follow, value_follows, "Unknown"),
> attributes_follow);
Because that is conditional on if (tree) while the variable is then used in a
subsequent switch conditional regardless of tree.
> - 'dissect_nfs_nfsstat4' doesn't exist in my version of the code. The patch
> changes 'dissect_nfs_nfsstat4' to 'dissect_nfs_status4'. How could this be?
Ah, there was an additional reference added in r48849 which was committed after
you posted your patch. I can fix that myself if necessary.
> - The "<field_name> duplicates PROTOABBREV of nfs" errors are true in the
> case
> of nfs.nfsstat3, nfs.nfsstat4, and hf_nfs_nfsstat all of which set
> PROTOABBREV to "Reply status" but there is nothing wrong with that.
> Shouldn't we just ignore those warnings (or remove that check from the
> script?) =)
I think it's just that nfs.stat3 makes more sense than nfs.nfsstat3 or
nfs.nfs.nfsstat3 etc. unless those names make more sense with reference to the
protocol (I don't know a whole lot about NFS internals to be honest). They're
not a big deal, I just thought they might be worth fixing while we're doing
cleanup.
You are receiving this mail because:
- You are watching all bug changes.