rsync-2.5.1 / socket.c - unititialized variable breaks build.

[Dave Dykstra]

It looks like Martin just put this fix into CVS because he turned on -Wall
on gcc and noticed a warning message.

- Dave Dykstra

On Mon, Jan 07, 2002 at 09:34:27PM -0600, John Malmberg wrote:
> Probably, but that just means that there was an addtional bug in the 
> original code.
> 
> At this phase, I am just trying to get rid of the compiler diagnostics 
> with out changing what the code is intending on doing.
> 
> Right now I am trying to make sense of what "char" declarations should 
> really be "unsigned char" or "void" before I submit the patches for them.
> 
> -John
> wb8tyw at qsl.network
> Personal Opinion Only
> 
> Cameron Simpson wrote:
> 
> > On Sun, Jan 06, 2002 at 10:56:50AM -0600, John Malmberg <wb8tyw at qsl.net> wrote:
> > | In socket.c, the variable error is unitilialzed.
> > | Also the variable length should be an unsigned int.
> > | 
> > | 
> > | EAGLE> type socket.gdiff
> > | --- ref_src:socket.c    Sat Jan  5 13:27:07 2002
> > | +++ lcl_src:socket.c    Sun Jan  6 08:40:51 2002
> > | @@ -1,3 +1,4 @@
> > | +/* Converted by prj_src:unix_c_to_vms_c.tpu AND prj_src:socket.tpu on 
> > | 6-JAN-20
> > | 02 08:40:51.01 OPENVMS_AXP */
> > |   /* -*- c-file-style: "linux" -*-
> > | 
> > |      Copyright (C) 1992-2001 by Andrew Tridgell <tridge at samba.org>
> > | @@ -105,7 +106,8 @@
> > |          bhints.ai_family = ai_family;
> > |          bhints.ai_socktype = ai_socktype;
> > |          bhints.ai_flags = AI_PASSIVE;
> > | -       if (getaddrinfo(bind_address, NULL, &bhints, &bres_all) == -1) {
> > | +       error = getaddrinfo(bind_address, NULL, &bhints, &bres_all);
> > | +       if (error == -1) {
> > 
> > Shouldn't that be "error != 0" ?
> >