Hi Alban! Alban browaeys [2005-06-02 19:51 +0200]: > > Yes, this can lead to all sorts of weirdnesses. I will add a test to > > pg_ctlcluster that will refuse to start the postmaster if the account > > is not valid to prevent such failures in the future. > > That s way better than my hack . I also though it would be great to > manage it at a higher level (less bloat then adding check in every part > of the code that deal with uid, better do it at the entry point). Though > i still have a comment ...
This is already fixed in arch head, BTW. > > if [ "`$PGPATH/pg_id -u`" = 0 ] > > > == would be posix complient. Though i am against the use for test ([]) when > testing a command output (instead of a variable). I am addicted to : > if `$PGPATH/pg_id -u` ; But this tests the exit code, not the output. Could be right, of course, but I did not check. > > I am used to always putting quote around shell variables though this > time i found it great you did not. I'm used to that, too. BTW, I did not write pg_ctl, that's upstream code. :-) I wrote the perl infrastructure around it (postgresql-common). > Maybe the check at pg_ctlcluster level is enough :) Yes, I agree. Have a nice day and thanks for the report! Martin -- Martin Pitt http://www.piware.de Ubuntu Developer http://www.ubuntulinux.org Debian Developer http://www.debian.org
signature.asc
Description: Digital signature
