Ethereal-users: Re: [Ethereal-users] libwiretap.a(toshiba.o) compile issue

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: Guy Harris <gharris@xxxxxxxxx>
Date: Sat, 4 Aug 2001 13:05:59 -0700
On Sat, Aug 04, 2001 at 09:41:55AM -0500, Nathan Neulinger wrote:
> I've experienced this many times before... It's because of Xfree86
> (possibly only 4.x)... It includes a libz

...which is a botch, on systems that have libz.

X11 has, as one of its many many compile-time options, "HasZlib", which
specifies whether the platform has libz; any OS that comes with libz
should, in its "config/cf/{platform}.cf" file, put in 

	#define HasZlib          YES

XFree86 3.3.6 doesn't have that in "linux.cf"; Linux distributors should
*NOT* just blithely assume that "linux.cf" applies unmodified to them,
as Not All Linux Distributions Are The Same, and the configuration file
should apply to the *entire operating system*, not just to the kernel
(although even the "but all Linuxes have the same kernel!" argument
against Linux being fragmented is bogus, as various distributions pick
up various patches that haven't had the Holy Penguin Pee, to use
Torvald's term, sprinkled on them).

The XFree86 CVS tree, as of a few seconds ago, *does* have "HasZlib" in
the "linux.cf" file; it also had it as of a while ago, before 4.1.0 came
out, so 4.1.0 probably has it, and all the 4.x releases might have it.

So the Slackware folks should be informed that they should add

	#define HasZlib          YES

to the "config/cf/linux.cf" file in the XFree86 3.3.x source before
building it, the fact that it would then not be Vanilla XFree86 Source
nonwithstanding (given a choice between vanilla code and code that
doesn't screw up, I tend to prefer the latter).