From: Jeff King <peff@peff.net>
Date: Wed, 10 Apr 2013 18:51:15 GMT
Subject: Re: segfault in git-remote-http
Message-ID: <20130410185114.GA18084@sigill.intra.peff.net>
In-Reply-To: <20130410090850.bacd762ad52eb3643ca99927@lavabit.com>

On Wed, Apr 10, 2013 at 09:08:50AM -0700, rh wrote:

> > which should show both program names. Git invokes git-remote-* based
> > on the URL you fed it. So if you are seeing a segfault in
> > git-remote-http, presumably you fed it an http URL (which may still
> > execute SSL code if it redirects to an https URL).
> 
> Here's the command I ran (from initial post):
> git clone https://github.com/bitcoin/bitcoin.git
> 
> This returns no error on the command line and produced the segfault
> reported by the kernel. git clone returns immediately.

It does correctly report a failed exit code. The lack of message is
because git assumes that the helper will produce a useful message before
dying, but obviously it doesn't.  There's already a patch[1] to fix this,
but it hasn't been merged yet.

As for why dmesg reports git-remote-http, I'm not sure. If you "strace
-f" the command, you can see that git is running git-remote-https. Why
the kernel chooses to report "git-remote-http", I don't know; you'd have
to look into how the kernel makes that decision. But I doubt it is
related to the reason for the segfault in the first place.

-Peff