Wade.Stuart at fallon
Sep 4, 2008, 4:26 PM
Post #3 of 3
(561 views)
Permalink
|
Bruce Keeler <bruce[at]drangle.com> wrote on 09/04/2008 01:26:11 PM:
> I'd go one step further, and have that code just automatically redact
> the values of any parameters matching something like /^pass(w(or)?d)?$/i
> by default. Logging passwords is just plain bad, there's almost never
> any valid reason to do it.
>
> I like to leave debug logging on even for production systems. Disk
> space isn't an issue, and it's extremely useful to be able to figure out
> exactly what a user was doing when they report a bug.
I do not like this, yuk. If this is considered a good idea and moves
forward please consider doing this only in Debug mode. If these are
getting generated any time besides Debug time (dumping raw params), then
the modules dropping the log lines should be sanitized. The auth modules
as far as I can tell do not dump the user/pass to log. Please don't make
assumptions about my log lines.
For instance we have at least two apps here that dump user:password pair
logs on failure to log in. These passwords are md5'ed for the log entry so
as we can tell if the user is trying different passwords, or the same
password over and over without compromising password secrecy.
-Wade
>
> Georg Oechsler wrote:
> > Hi,
> >
> > I need to filter the log output of an intranet app, because it
> > authenticates users against our LDAP server and it is not desired to
> > have the LDAP passwords of my colleagues in the logs - even if
> > debugging is on for only a few minutes to track down a problem.
> >
> > I found this topic was on the list a few years ago:
> >
http://lists.scsys.co.uk/pipermail/catalyst-dev/2006-January/000077.html
> >
> >
> > The proposed solution was to suppress output with $c->log->abort(1).
> > Unfortunately this also supresses the username, which makes searching
> > logs quite painful.
> >
> > So i started to hack a solution for this problem. First I overode
> > prepare_body() in MyApp.pm which works fine as long as nobody touches
> > it in Catalyst.pm. Then I tried to write a plugin, but it's same
> > problem - I have to touch the plugin everytime the logic of
> > prepare_body() changes.
> >
> > Finally I encapsulated part of the code which writes the debug
> > information to the function debug_parameters().
> > In Catalyst.pm it does what it did before, in MyApp.pm it prints out
> > whatever I like - especially no passwords.
> >
> > I attached a patch against Catalyst.pm Version 5.7014.
> >
> > What do you think of this?
> >
> > Georg
> >
> >
> >
> >
> >
> >
> >
> >
------------------------------------------------------------------------
> >
> > _______________________________________________
> > Catalyst-dev mailing list
> > Catalyst-dev[at]lists.scsys.co.uk
> > http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst-dev
>
>
> _______________________________________________
> Catalyst-dev mailing list
> Catalyst-dev[at]lists.scsys.co.uk
> http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst-dev
_______________________________________________
Catalyst-dev mailing list
Catalyst-dev[at]lists.scsys.co.uk
http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst-dev
|
debug_parameters.patch
smime.p7s
