Update of bug #2718 (project savane):
Priority: A - Later => C - Normal
Status: None => Fixed
Assigned to: None => beuc
Open/Closed: Open => Closed
Release: => 1.0.7
Planned Release: => 1.0.8
_______________________________________________________
Follow-up Comment #3:
> I used my script to update my test install and Savannah - so I
> tested it :)
I should have guessed that ...
> If you test your script however, you'll see that system
> returns 0 (the process exit code, not a boolean), which will
> trigger the die() part ;)
>
> Another way is to use "and die()", but that's not very clear,
> hence the C-style "if (sthing != 0) die"
Alright, my fault. I was so used to the syntax from the "open" calls.
> Suppressing -u and -p in the mysql call falls back to the
> username and/or info in ~/.my.cnf. Using only -p forces
> requesting password, even if it is in ~/.my.cnf. Usually the
> script is ran by somebody used to accessing the database
> manually, hence it's likely (s)he uses a .my.cnf file. In my
> case user root was used, ie the one with the privs to drop and
> delete the database (and create new tables).
>
> Btw, something I just realize is that using -p $PASS is
> insecure because the pass will show up in the process list (ps
> aux...).
Both are valid points. Please commit your version of convert_database.pl to
the CVS, and thanks for pointing out those errors.
One tiny remark, though: I'd replace the second comment "# drop current
database" with "# re-create current database".
_______________________________________________________
Reply to this item at:
<http://gna.org/bugs/?func=detailitem&item_id=2718>
_______________________________________________
Message sent via/by Gna!
http://gna.org/
_______________________________________________
Savane-dev mailing list
[email protected]
https://mail.gna.org/listinfo/savane-dev
