- From: Mike West <mkwst@google.com>
- Date: Tue, 3 Jul 2012 14:23:08 -0500
- To: Odin Hørthe Omdal <odinho@opera.com>
- Cc: public-webappsec@w3.org
- Message-ID: <CAKXHy=eqN=Kw5aQbP4QaOT+pCwrG48jOa0B9ohFybnGFywjsPQ@mail.gmail.com>
Note that WebKit only started complying with CSP 1.0's spec on this point about two weeks ago (http://trac.webkit.org/changeset/120540). Prior to that point, it was (silently) erroring off on the whole source. Given that history, we might have success at preparing the way for granularity in 1.1 by adding a warning to 1.0 implementations now, noting that the path component is being ignored. That would be lighter-weight than a version component. -mike -- Mike West <mkwst@google.com>, Developer Advocate Google Germany GmbH, Dienerstrasse 12, 80331 München, Germany Google+: https://mkw.st/+, Twitter: @mikewest, Cell: +49 162 10 255 91 On Mon, Jul 2, 2012 at 6:45 AM, Odin Hørthe Omdal <odinho@opera.com> wrote: > On Fri, 22 Jun 2012 11:31:41 +0200, Mike West <mkwst@google.com> wrote: > >> One of the proposals for CSP 1.1 is additional granularity in source >> paths (http://www.w3.org/Security/**wiki/Content_Security_Policy#** >> Proposals_for_Version_1.1<http://www.w3.org/Security/wiki/Content_Security_Policy#Proposals_for_Version_1.1>). >> I think this additional granularity is well worth perusing >> > > I think so too. There's many places in CSP that I think it's a bit too > granular and rather too complex IMHO, but this case seems a quite common > way to give some additional security to smaller sites. > > In fact, it was also the first thing that came up when I talked with > hackers making a small locally hosted image gallery software. > > You can work around it by having a userfiles domain, but it would > complicate the setup procedure immensely. > > > > The problem with how the spec is doing things now (throwing away path > component) is that sites using CSP (1.0) will no doubt have errors. They'll > write script-src: http://my-site.com/js/ and use scripts from js, except > for that one time they use one on /my-demo/js.js and it works anyway so > they actually don't think about it. > > So if CSP 1.0 is allowed to live a long time in a browser, the behavior we > have now might actually be mandatory for site-compat. > > -- > Odin Hørthe Omdal (Velmont/odinho) · Core, Opera Software, > http://opera.com > >
Received on Tuesday, 3 July 2012 19:24:01 UTC