On 14/03/19 13:16, Peter Maydell wrote: > Mmm, that makes sense. So in this particular case, do we > want to try to avoid doing an unbounded allocation based > on whatever rubbish the user passed us in the environment, > or do we say "this particular case is OK" and mark it > as a false-positive ? > > Cc'ing Gerd since the Coverity issues in question are > in the audio code (in get_str() and in > audio_handle_legacy_opts()).
I think the solution for that is deprecating the old legacy options and getting rid of them in one year... Paolo