335457 - [FIX]Make URI vs originalURI code consistent in document and docshell

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Boris Zbarsky [:bzbarsky]]

Right now, documents and docshells do different things wrt originalURI vs URI for incoming data.  We should really fix that.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: Proposed patch

I'd like this on the 1.8.1 branch too; Ben's feed stuff is broken without this.

Note that this affects behavior of two cases:

  chrome/resource/about originalURI and LOAD_REPLACE set.  This currently does
     not happen.
  some other originalURI and LOAD_REPLACE not set.  Since current HTTP redirects
     do set LOAD_REPLACE, we'd ok.

In any case, this makes documents sorta follow the nsIChannel docs on what LOAD_REPLACE means.

Comment 2

[Ben Goodger (use ben at mozilla dot org for email)]

Marking blocking-1.8.1, as long as this is in the critical path for feed handling

Comment 3

[Daniel Veditz [:dveditz]]

Comment on attachment 219818 [details] [diff] [review]
Proposed patch

r=dveditz

Should regression-check bug 251368

Comment 4

[Boris Zbarsky [:bzbarsky]]

Did that.  It looks good.

Fixed on trunk.

Comment 5

[Dave Townsend [:mossop]]

A local backout shows that this caused bug 337339

Comment 6

[Mike Schroepfer]

Bz - do we need both this and bug 337339 for 1.8.1?

Comment 7

[Boris Zbarsky [:bzbarsky]]

Hmm...  I believe we need this for bug 325085, right?  If so, we probably do need this on the 1.8.1 branch, and then we need a fix for bug 337339 there.  Unfortunately, the patch for bug 337339 can't be used on the branch as-is; we'd need to either take an entirely different approach or at the very least add an nsIURIFixup_MOZILLA_1_8_BRANCH if we decide removing keyword: on the branch is ok...

Comment 8

[Mike Schroepfer]

Holding the approval on this until we get a branch patch for the regression bug 337339 - would be great to land that and this together.

Comment 9

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Comment on attachment 219818 [details] [diff] [review]
Proposed patch

Clearing nomination so we don't go over it again at each triage meeting.  Please renominate once it's ready (see previous comment).

Comment 10

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 219818 [details] [diff] [review]
Proposed patch

OK, the regression fix landed.  So this is good to go for branch.

Comment 11

[Boris Zbarsky [:bzbarsky]]

Fixed on 1.8 branch.