read error' when tripping on first_byte_timeout, and 'http read error'
when tripping on between_bytes_timeout (used to be 'http first read
error' in both cases). This now matches current master.

This looks like a bit of forgotten debugging, left by mistake.

Fixes	#1312

See Also:	CVE-2013-4090

Fixes #1285

Fixes: #1076

Some of them are rather pedantic:  "NULL" vs "(char*)0" -- Really ?

Remove a bunch of ifdef code and fix up proper configure check for
libedit.

Fixes #1277

The test case for this doesn't trigger, but the failing VCL (when
given in a file with -f) looks like:

director squid round-robin {
	 { .backend = { .host = "127.0.0.1"; .port = "3131"; } }
}

director dnsdir dns {
	 .list = {
	       "201.7.184.0"/32;
	 }
}

sub vcl_recv {
    set req.backend = dnsdir;
}

Fixes #1295

If there are multiple matches for a given VSL tag (such as "Hit"
matching both Hit and HitPass), prefer the one that is an exact match.

Fixes #1288

If we don't add -c and -b in there, varnishlog will print all
transactions that aren't associated with a session.  That is likely to
not be what the user wants.

Fixes: #1071

poll can return EINTR, so we need to handle that.

Fixes: #1166

parameter to preserve the arguments as seen e.g. ps.

Fixes: #1238

this information is no longer available.

Fixes #1228

Fixed by: Federico G. Schwindt
Fixes: #1216

set req.http.Cookie = req.http.Cookie + " " + req.http.Cookie-sess;

do the obvious thing, even if req.http.Cookie does not exist.

The underlying issue was a badly though through overloading of
the NULL value to mean "Unset".  Now it has its own magic marker.

Fixes #1218

ban() was previously not in bold on the man page. Turn it into two
separate entries with separate descritions to fix that.

the strings we get from the libvarnishapi functions are not
necessarily null-terminated, so avoid using str* functions on them.

Thanks to lampe for the patch.

Fixes #1115