David Levine <levinedl@acm.org> writes:
Ken wrote:
I suggest we simply remove closefds() completely.
Great idea!
Makes sense from here. You could back-fill use of CLOEXEC anywhere
it seemed important.
Alright, done.
In the "weird things I discovered while doing this" ... I saw in slocal
that when mail is delivered to a process via a pipe, right after fork()
and right before exec(), it uses dup2() to duplicate the file descriptor
pointing to the input message down to 0 (expected), opens /dev/null and
redirects standard output and standard error to it (expected), but ALSO
uses dup2() to make an additional copy of the input file descriptor to
descriptor 3 (!). Does anyone know why? It looks like it has always
done this.
--Ken
--
nmh-workers
https://lists.nongnu.org/mailman/listinfo/nmh-workers