On Sat, Aug 7, 2021 at 10:11 PM Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> On 8/7/21 11:42 AM, Warner Losh wrote:
> > + path = g_strdup(p);
> > + if (path == NULL) {
>
> Only returns null when the input is null, which you've already eliminated.
>
> > +static bool find_in_path(char *path, const char *filename, char *retpath,
> > + size_t rpsize)
> > +{
> > + const char *d;
> > +
> > + while ((d = strsep(&path, ":")) != NULL) {
> > + if (*d == '\0') {
> > + d = ".";
> > + }
> > + if (snprintf(retpath, rpsize, "%s/%s", d, filename) >= (int)rpsize) {
> > + continue;
> > + }
> > + if (is_there((const char *)retpath)) {
> > + return true;
> > + }
> > + }
> > + return false;
>
> Hmm. Fixed size retpath buffer isn't ideal.
> Any reason not to use g_find_program_in_path?
>
g_find_program_in_path may work well here, as well...
Quite well: 1 file changed, 5 insertions(+), 42 deletions(-) so a nice
reduction in code when the unnecessary stuff is removed. I'll have
that in v2 of the series.
> I note that we don't search the path at all in linux-user/.
>
IIRC imgact_binmisc will have the resolved path but preserve argv as
it should have been were it not emulated, so we have to re-evaluate
the PATH search here because we try to be faithful to the context.
I just looked at the imgact_binmisc code, and that's what it does, so
that sounds reasonable to me as well.
Warner