At Fri, 17 Dec 2021 02:42:25 +0000, "kuroda.hayato@fujitsu.com" <kuroda.hayato@fujitsu.com> wrote in
> Dear Fujii-san,
>
> Sorry I forgot replying your messages.
>
> > >> if (strcmp(keywords[i], "application_name") == 0 &&
> > >> values[i] != NULL && *(values[i]) != '\0')
> > >
> > > I'm not sure but do we have a case that values[i] becomes NULL
> > > even if keywords[i] is "application_name"?
> >
> > No for now, I guess. But isn't it safer to check that, too? I also could not find strong
> > reason why that check should be dropped. But you'd like to drop that?
>
> No, I agreed the new checking. I'm just afraid of my code missing.
FWIW, I don't understand why we care of the case where the function
itself changes its mind to set values[i] to null while we ignore the
possibility that some module at remote is modified so that some global
variables to be NULL. I don't mind wheter we care for NULLs or not
but I think we should be consistent at least in a single patch.
regards.
--
Kyotaro Horiguchi
NTT Open Source Software Center