On Tue, Sep 19, 2023 at 09:23:47PM +0200, Christophe JAILLET wrote: > Le 19/09/2023 à 21:02, Kent Overstreet a écrit : > > On Tue, Sep 19, 2023 at 09:17:27AM -0400, Brian Foster wrote: > > > On Sat, Sep 16, 2023 at 09:30:19AM +0200, Christophe JAILLET wrote: > > > > snprintf() and scnprintf() are the same, except for the returned value. > > > > When this value is not used, it is more logical to use snprintf() which is > > > > slightly simpler. > > > > > > > > Signed-off-by: Christophe JAILLET <christophe.jaillet@xxxxxxxxxx> > > > > --- > > > > > > Seems reasonable: > > > > > > Reviewed-by: Brian Foster <bfoster@xxxxxxxxxx> > > > > No, let's stay with scnprintf as the default - snprintf should be > > deprecated except for when its return value is actually needed, using it > > incorrectly has been a source of buffer overruns in the past. > > > > Ok, I was not aware of it. > > In this case, there are also some s/snprintf/scnprintf/ opportunities in > fs/bcachefs > > Does it make sense to update them or is it too low value changes? Not terribly important - long term, I want to depracate both snprintf and scnprintf and convert everything to printbufs.
