-----Original Message-----
From: "Andreas Pflug"<[EMAIL PROTECTED]>
Sent: 01/05/06 11:46:24
To: "[EMAIL PROTECTED]"<[EMAIL PROTECTED]>
Cc: "[email protected]"<[email protected]>
Subject: Re: [pgadmin-hackers] SVN Commit by dpage: r5106 - in trunk/pgadmin3:

> Is that a good idea to do this in a private md5 implementation? libpq is 
> going to/already exporting a method for this, which will make us less 
> sensitive to backend changes.

I considered this carefully and chose that method for a few reasons.

- Any change (say, to sha1) is a major compatibility issue for any libpq client 
- it wouldn't be taken lightly, we couldn't fail to miss the change, and it'll 
be mainly a cut n paste job to fix.

- Using libpq to do it requires configure time checks for the correct libpq on 
*nix, and runtime checking on Win32. I couldn't find a sensibly clean way to do 
the latter.

- Using libpq essentially forces ppl to upgrade to the latest client to benefit 
from the extra security.

All in all, a private implementation seemed the least painful, and secure 
option.

/D


-----Unmodified Original Message-----
[EMAIL PROTECTED] wrote:
> Author: dpage
> 
> Date: 2006-04-30 21:13:00 +0100 (Sun, 30 Apr 2006)
> 
> New Revision: 5106
> 
> Revision summary: 
> http://svn.pgadmin.org/cgi-bin/viewcvs.cgi/?rev=5106&view=rev
> 
> Log:
> Pre-encrypt passwords before sending them down the wire or displaying them in 
> SQL statements.

Is that a good idea to do this in a private md5 implementation? libpq is 
going to/already exporting a method for this, which will make us less 
sensitive to backend changes.

Regards,
Andreas

---------------------------(end of broadcast)---------------------------
TIP 5: don't forget to increase your free space map settings

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster

Reply via email to