Re: [PATCH] http: fix segfault in handle_curl_result

From: Jeff King <peff@peff.net>
Date: Oct 12, 2012, 07:30 UTC
Message-ID: <20121012073053.GC17026@sigill.intra.peff.net>
In-Reply-To: <20121012062249.GB17026@sigill.intra.peff.net>

On Fri, Oct 12, 2012 at 02:22:49AM -0400, Jeff King wrote:

Show 15 quoted lines

> We should probably not be passing the slot to handle_curl_results at
> all, since it may have already been reused and is not safe to read. The
> only thing we do with it is to set up any new auth information in the
> curl handle.  This doesn't suffer the same problem because a reused slot
> will always have a curl handle. However, it means we may be setting the
> auth information for a random handle. Which is OK, since all handles use
> the same auth information anyway.  But it should also be pointless,
> because since dfa1725 (fix http auth with multiple curl handles) we
> always refresh the curl handle's auth information whenever we get an
> active slot.
> 
> However, I'm leaving that out of this patch.  Commit 8809703 was
> supposed to be a refactor with zero behavior changes, but it regressed.
> This fixes the regression by behaving exactly as we did beforehand. We
> can build the other thing on top.

I took a look at this, and indeed, it breaks existing code. But the broken code is wrong and is easy to fix. So here is a series to do this, on top of the one I am responding to.

These ones shouldn't have any functional impact, but do clean up the code. I'd recommend the regression fix I already posted for maint and 1.8.0, and leave these to cook for post-1.8.0.

  [1/2]: remote-curl: do not call run_slot repeatedly
  [2/2]: http: do not set up curl auth after a 401

-Peff