Hey Stanton, Could you upload the patch to http://codereview.appspot.com/ and CC me and/or Jacobo, w/ [email protected] on there too?
-j On Mon, Apr 4, 2011 at 3:12 PM, Stanton Sievers <[email protected]> wrote: > Hi John and Jacobo, > > Letting the browser handle the scema-less URL makes sense. I've played > around with this a bit and I think I have a solution. I've attached a patch. > > These changes will break unlockedDomain URLs such as "www.example.com" and > "www.example.com:8080" when they used to work. These are now broken > because I moved the setAuthority() call outside of the conditional in in > DefaultIframeUriManager. Based on RFC 3986, a URI such as "www.example.com" > is parsed as the Path not the Authority and one such as " > www.example.com:8080" is parsed with "www.example.com" as the Scheme and > "8080" as the Scheme-specific Part. The setAuthority() call in the case of a > blank schema worked around that, although probably incorrectly. Unlocked > domain URLs would now require either a Schema or "//" prepended to the URL > in order to unambiguously be parsed as the Authority. > > If this is going to cause too much turbulence, I could do some further > checking as was done before. > > Thoughts? > > Best regards, > -Stanton > > *(See attached file: DefaultIframeUriManagerPatch.txt)* > > [image: Inactive hide details for John Hjelmstad ---04/04/2011 > 13:51:41---@Stanton, the way Jacobo's describing is how we use it at Goo]John > Hjelmstad ---04/04/2011 13:51:41---@Stanton, the way Jacobo's describing > is how we use it at Google -- we just rely on the browser to d > > From: John Hjelmstad <[email protected]> > To: [email protected], > Cc: Jacobo Tarrio <[email protected]>, Stanton Sievers/Westford/IBM@Lotus > Date: 04/04/2011 13:51 > Subject: Re: DefaultIframeUriManager for locked domains > ------------------------------ > > > > @Stanton, the way Jacobo's describing is how we use it at Google -- we just > rely on the browser to do the resolution. > > Adding scheme to the config doesn't work for us since we have containers > (multiple actually) that run in both HTTP and HTTPS mode. We'd need a > substitution var to make a single config work, with substitution provided > by > an injected value. > > --j > > On Fri, Apr 1, 2011 at 9:33 AM, Jacobo Tarrio <[email protected]> wrote: > > > On Fri, Apr 1, 2011 at 5:07 AM, Stanton Sievers <[email protected]> > > wrote: > > > > > > > Regarding blank schemes, does that mean that in this particular case, > the > > > iframeurl returned in the metadata request would be schema-relative and > > it > > > would be up to the container to add that information before rendering > the > > > iframeurl? > > > > > > > Or not; the browser will see the schema-less URL and use the schema of > the > > page that contains it. > > > > -- > > Jacobo Tarrío | http://jacobo.tarrio.org/ > > > > >
