Comment # 3
on bug 7471
from Michael Mann
(In reply to comment #2)
> Beware of throwing out UATs for dissectors not found! Imagine this: For some
> reason my plugin cannot be loaded (network disk error, file handling
> mistake, whatever reason). Now if that would nuke my carefully compiled UAT,
> with all my lab settings carefully crafted for over a month, I would be very
> unhappy.
So you think an exception should be made for "protocol" fields in that they
don't need to be validated and the onus is on the dissector using the UAT?
What about the other field types? I like that a dissector doesn't have to keep
validating UAT fields during dissection.
Do we need two validation procedures - one for GUI and one for loading (at
startup)? Right now the same validation function is done in the GUI (check
that dissector exists) as startup. Technically I don't think the UAT entry is
removed from the (hidden) file its stored in unless the UAT is selected in the
GUI, its just removed from the list that goes into memory (which the GUI uses).
Should the GUI use the "raw" file (and no field validation when loading)? If
that's the case, I don't know how you would let the user know there are errors
unless a specific UAT entry is chosen. And then what would you do with 2
invalid entries?
You are receiving this mail because:
- You are watching all bug changes.