[PATCH 8 of 8] http: reuse authentication info after the first failed request (issue3567)

[Matt Mackall]

On Tue, 2014-01-21 at 01:16 +0100, Mads Kiilerich wrote:
> On 01/21/2014 12:19 AM, Matt Mackall wrote:
> > On Mon, 2014-01-20 at 01:32 +0100, Mads Kiilerich wrote:
> >> # HG changeset patch
> >> # User Stéphane Klein <contact at stephane-klein.info>
> >> # Date 1387547765 -3600
> >> #      Fri Dec 20 14:56:05 2013 +0100
> >> # Node ID 1748dd9992a0f2eea93eb8afb1ed2804373dbf55
> >> # Parent  5ec933e1841d7ea04fee58f00b0e9f9929da2ca2
> >> http: reuse authentication info after the first failed request (issue3567)
> > This is tricky. This is clearly a bugfix, but there's some question
> > about whether it's the right fix, it's at the end of a series of
> > miscellaneous changes that aren't all stable-appropriate, and arrived
> > after the nominal freeze date. It looks like I can't take it by itself
> > as its test will break.
> >
> > So, if you're confident this is the way to go, please pull just the fix
> > portion of this out and submit for stable, and we'll revisit the other
> > bits for post-2.9.
> >
> > I've queued patches 3 and 5, which looked sufficiently bug-fixy and
> > orthogonal to go in on their own, thanks.
> 
> 2 is necessary to let the tests pass on Fedora. It can be taken without 
> the 1 that made it possible to find the problem.
> 
> 6 fixes existing test failure with Python 2.4, 2.4.1 and 2.4.2. The 
> 2.4.2 on the build farm must be a patched version.

It is.

> 8 is an augmented version of Stéphane's original patch, augmented with 
> some test coverage. If you want it without 7 then just ignore the test 
> changes ... or take Stéphane's version.

Standard operating procedure applies: nothing is happening here without
a resend, because everything else has long since been deleted from my
inbox.

-- 
Mathematics is the supreme nostalgia of our time.