Re: BUG #15120: use of getcwd(3)/chdir(2) during path resolution (exec.c)

On Sat, Mar 17, 2018 at 12:28:53PM -0400, Tom Lane wrote:
> PG Bug reporting form <noreply(at)postgresql(dot)org> writes:
> > In the patch below, path resolution is skipped if getcwd returns EACCES.
>
> ... I find that quite an unacceptable answer.

The patch was just a POC, yes. I was out of better ideas and not really
qualified to rewrite find_my_exec/resolve_symlinks without breaking path
resolution for some other supported platform.

> The idea of pre-checking to see if the initial path is already absolute
> seems safe enough, but I'm not sure how much of the use-case it'd cover.
> I think your example of "sudo /usr/bin/pg_ctl" is pretty artificial;
> who'd bother spelling that out?

rc-scripts on some BSDs spell it out:

$ uname
NetBSD

$ egrep 'command=|doit=' /etc/rc.d/pgsql
command="/usr/pkg/bin/pg_ctl"
doit="/usr/bin/su -m ${pgsql_user} -c '${command} init ${command_args}'"
doit="/usr/bin/su -m ${pgsql_user} -c '${command} start ${command_args}'"
doit="/usr/bin/su -m ${pgsql_user} -c '${command} restart ${command_args}'"
doit="/usr/bin/su -m ${pgsql_user} -c '${command} stop ${command_args}'"
doit="/usr/bin/su -m ${pgsql_user} -c '${command} reload ${command_args}'"

But you are right, testing for an absolute path is not a complete
solution either. Running "sudo pg_ctl" will still fail.