[Jeff Morriss <jeff.morriss.ws@xxxxxxxxx>]

On Mon, Nov 12, 2018 at 4:42 AM Dario Lombardo <lomato@xxxxxxxxx> wrote:

On Mon, Nov 12, 2018 at 9:40 AM Antoine d'Otreppe <a.dotreppe@xxxxxxxxxx> wrote:

2. I used the fuzz-test.sh, looks nice, but how many passes should I run to have adequate testing?

 

That's an interesting question. I don't recall how many passes I used in the past for fuzzing my dissectors. Moreover no default number of passes are present in the script. I guess once we've found the answer we could set it in the fuzzing script.

Anyone more experienced in dissectors have an answer?

There's no magic number, of course.  It depends on how many on how many captures your testing and how big they are.  Personally I used to run it at least for half an hour or so while I was off doing other things (like beautifying the code, committing it, etc.); for a brand new dissector I might go longer.  Or I'd just leave it running while I went off and ate lunch or dinner or whatever.