460124 - remove workaround for principal hashtable inconsistency

Status: RESOLVED FIXED

(Firefox :: General, defect)

Description

[Ben Newman (:bnewman) (:benjamn)]

User-Agent:       Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.0.1) Gecko/2008070206 Firefox/3.0.1
Build Identifier: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.0.1) Gecko/2008070206 Firefox/3.0.1

With bug #454850 fixed and pushed, this workaround (in nsScriptSecurityManager.cpp) becomes unnecessary:

mPrincipals.Get(principal, getter_AddRefs(fromTable));
if (!fromTable)
{
    //-- Check to see if we have a more general principal

    // XXXbz if only GetOrigin returned a URI!  Or better yet if the
    // HashKey function on principals were smarter.  As it is, we can
    // have cases where two principals will have different hashkeys but
    // test equal via KeyEquals, which is absolutely silly.  That's
    // what we're working around here.
    nsXPIDLCString originUrl;
    rv = principal->GetOrigin(getter_Copies(originUrl));
    if (NS_FAILED(rv)) return rv;
    nsCOMPtr<nsIURI> newURI;
    rv = NS_NewURI(getter_AddRefs(newURI), originUrl, nsnull, sIOService);
    if (NS_FAILED(rv)) return rv;
    nsCOMPtr<nsIPrincipal> principal2;
    rv = CreateCodebasePrincipal(newURI, getter_AddRefs(principal2));
    if (NS_FAILED(rv)) return rv;
    mPrincipals.Get(principal2, getter_AddRefs(fromTable));
}

That is, the first result from mPrincipals.Get can now be trusted.

Reproducible: Didn't try

Comment 1

[Ben Newman (:bnewman) (:benjamn)]

Attachment: proposed patch

Tempted to unify the similar parts of DoGetCertificatePrincipal with GetCodebasePrincipal, but that's not really at issue here.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 343297 [details] [diff] [review]
proposed patch

r+sr=bzbarsky

If you do want to do that unification, that sounds pretty good to me!

Comment 3

[Boris Zbarsky [:bzbarsky]]

Pushed changeset 1b853ea8e418.