Re: pkg/49490: "hg clone" does not work on big endian machines

[christos%astron.com@localhost (Christos Zoulas)]

In article <20150110201501.26744A65BA%mollari.NetBSD.org@localhost>,
Martin Husemann  <gnats-bugs%NetBSD.org@localhost> wrote:
>The following reply was made to PR pkg/49490; it has been noted by GNATS.
>
>From: Martin Husemann <martin%duskware.de@localhost>
>To: gnats-bugs%NetBSD.org@localhost
>Cc: 
>Subject: Re: pkg/49490: "hg clone" does not work on big endian machines
>Date: Sat, 10 Jan 2015 21:14:40 +0100
>
> I traced the network traffic while taking a ktrace at the same time,
> all running with a slightly modified py-mercurial, that would invoke
> the python debugger once the problem happens - and also doing
> hg --debug clone ...
> 
> If needed I can make the (significantly larger) full traces available, the
> relevant tail of both is here:
> 
> http://www.netbsd.org/~martin/hg-clone-pcap.tar.bz2 (2.3 MB)
> 
> It seems the actual 0 byte return comes from a recvfrom(), the socket
> connected to the http port seems to be fd 4:
> 
>  12647      1 python2.7 CALL  recvfrom(4,0xfffffffff60ae024,0x6a12,0,0,0)
>  12647      1 python2.7 MISC  msghdr: [name=0x0, namelen=0,
>iov=0x1b2915cc0, iovlen=1, control=0x0, controllen=0, flags=0]
>  12647      1 python2.7 GIO   fd 4 read 0 bytes
>        ""
>  12647      1 python2.7 RET   recvfrom 0
>  12647      1 python2.7 CALL  recvfrom(4,0xfffffffff60ae024,0x6a12,0,0,0)
>  12647      1 python2.7 MISC  msghdr: [name=0x0, namelen=0,
>iov=0x1b2915cc0, iovlen=1, control=0x0, controllen=0, flags=0]
>  12647      1 python2.7 GIO   fd 4 read 0 bytes
>        ""
>  12647      1 python2.7 RET   recvfrom 0
> 
> [..]
> and this imediately leeds to:
> 
>  12647      1 python2.7 CALL  write(2,0xfffffffff64f16e4,0x13)
>  12647      1 python2.7 GIO   fd 2 wrote 19 bytes
>        "transaction abort!\n"

What happens on the sender side?

christos