Re: Security: read vs. write

Eliotte,

    Thanks so much for the comment. Working Group member Jeremy Carroll
believes this would address your comment. The particular operation we
had in mind was from XSLT2: xsl:result-document. Perhaps we should make
this more explicit.

The rewrite of this section was motivated by implementer feedback.
Particularly concerning test security in

   http://jena.sourceforge.net/test/grddl/

which, with a little imagination, could be modified so that malicious 
code took control of an overly trusting machine (by writing 
appropriately to a key OS file). We are currently working on drafting a
more complete test-suite for GRDDL. Do you think this response addresses
your comment by itself, or
should we add a test for this directly to the test suite?


Elliotte Harold wrote:
>
> In section 8 I find:
>
> In particular, operations to read or write URLs are more safely
> executed with the privileges associated with an untrusted party,
> rather than the current user.
>
> I'm not sure what you're considering here with respect to the write
> half of this pair. Standard XSLT never writes any URL, and I wouldn't
> expect GRDDL to as a general rule. In other words, XSLT only GETs.
> never POSTs or PUTs.
>
> There are extensions to do this but you warn against them separately.
>
>


-- 
		-harry

Harry Halpin,  University of Edinburgh 
http://www.ibiblio.org/hhalpin 6B522426