Re: POC for a function trust mechanism - Mailing list pgsql-hackers

From Nico Williams
Subject Re: POC for a function trust mechanism
Date
Msg-id 20180809195656.GC5695@localhost
Whole thread Raw
In response to POC for a function trust mechanism  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Wed, Aug 08, 2018 at 01:15:38PM -0400, Tom Lane wrote:
> This is sort of a counter-proposal to Noah's discussion of search path
> security checking in <20180805080441.GH1688868@rfd.leadboat.com>.
> (There's no technical reason we couldn't do both things, but I think
> this'd be more useful to most people.)

So, this is why I always fully-qualify all references to functions,
tables, etc.  I also always set a search_path on each function just in
case I accidentally leave a non-fully-qualified symbol.

I would like to have a way to request that all non-fully-qualified
symbols be resolved at function create/replace time and that the
resolution results be made permanent for the function.  If I have
several schemas in a search_path at function definition time, this would
not allow me to move dependencies around without replacing the
dependents -- that's OK for me.

Nico
-- 


pgsql-hackers by date:

Previous
From: Nico Williams
Date:
Subject: Re: [FEATURE REQUEST] Encrypted indexes over encrypted data
Next
From: Alvaro Herrera
Date:
Subject: Re: Documentaion fix.