On 11/25/06, Woodchuck <[EMAIL PROTECTED]> wrote:
> On Fri, 24 Nov 2006, Bulk Mail wrote:
>
> > On Thursday 23 November 2006 09:15, Bulk Mail wrote:
> > > Hi,
> > >
> > > I need to run etherape but do not have X installed, what's a good way of
> > > adding the two?
> >
> > OK, having gotten X up, etherape installed, I'm getting a complaint that "No
> > nameservers defined. I've tested the local dns every way but Sunday, and it
> > all seem to work just fine.
> >
> > I got a local LAN under RFC1918. I can do both forward and reverse lookups
> > on
> > local and external addresses. Both named-checkconf and named-checkzone
> > passes
> > fine.
> >
> > I got caching on and master of my third level subdomain (to separate from
> > the
> > ISP hosted 2nd level domain).
> >
> > Obviously etherape is trying to do some lookup and fails, but I've got no
> > idea
> > where... If I start etherape without nameresolution it works, so it sure
> > looks like a dns problem. :(
> >
> > (Running OBSD 3.9)
>
> Well, hours later...
>
> I built this etherape from scratch and its first invocation showed me
> this:
>
> [EMAIL PROTECTED] root]# etherape
>
> (EtherApe:4707): libglade-WARNING **: Unknown GtkToolbar child property:
> sensitive
>
> (EtherApe:4707): libglade-WARNING **: Unknown GtkToolbar child property:
> sensitive
>
> (EtherApe:4707): GLib-GObject-WARNING **: g_object_set_property: object class
> `GtkTable' has no property named `orientation'
> No nameservers defined.
> [EMAIL PROTECTED] root]#
>
> I suppose that duplicates your problem.
>
> As far as I'm concerned this is typical of gnome shit and Linux software
> in general and is one very good reason not to run it. Cruft. Amateur
> programming. Undebugged portability. Huge libraries. Endless disk
> space. Bonobos. Pangos. Druids. And the shit is still shit, and
> doesn't work. At least this one isn't written in c-pus-pus.
>
> So I started it -nonum, and it opened an absolutely unreadable display,
> with various names in one column on the left (illegible colors, fixed
> size, names are truncated -- AMATEUR NIGHT). On the right is some
> ever-changeing goofball display with circles and triangles, in strange
> colors. Clicking on help reveals the name of the wanker who wrote this
> stuff five years ago, alias "Juan Toledo", and nothing else.
>
> whilst running it shits forth meaningless yet scary error messages like:
>
> (EtherApe:3764): GLib-GObject-WARNING **: g_object_set_property: object class
> `GtkTable' has no property named `orientation'
>
> (EtherApe:3764): Gtk-CRITICAL **: gtk_table_resize: assertion `n_cols > 0 &&
> n_cols < 65536' failed
>
> (EtherApe:3764): Gtk-CRITICAL **: gtk_table_resize: assertion `n_cols > 0 &&
> n_cols < 65536' failed
>
> (EtherApe:3764): Gtk-CRITICAL **: gtk_table_resize: assertion `n_cols > 0 &&
> n_cols < 65536' failed
>
> (EtherApe:3764): Gtk-CRITICAL **: gtk_table_resize: assertion `n_cols > 0 &&
> n_cols < 65536' failed
>
> (EtherApe:3764): Gtk-CRITICAL **: gtk_table_resize: assertion `n_cols > 0 &&
> n_cols < 65536' failed
>
> (EtherApe:3764): Gtk-CRITICAL **: gtk_table_resize: assertion `n_cols > 0 &&
> n_cols < 65536' failed
>
> (EtherApe:3764): Gtk-CRITICAL **: gtk_table_resize: assertion `n_cols > 0 &&
> n_cols < 65536' failed
>
> which means that the programming is unfinished and ships without
> documentation (like all Linux projects) and contains active "assert"
> macros in it. Maybe these are in the gtk bugware.
>
> When activated -noname with all else as default, this pig consumes
> 32% of the CPU on a 1GHz Pentium III, displaying on another machine's
> Xserver. Typical Linuxware.
>
> the man page is a joke, and I'll be double-damned if I'll start delving
> in the source code of this wankerware.
>
> Well, I'm double-damned. Here's a typical comment from its main.c
>
> /* TODO Besides the fact that this probably makes little sense nowadays
> * (at least for node color) it probably leads to a segfault
> * See how it is done for filter, for instance */
>
> Bah.
>
> OK, the error message comes from "dns.c"
>
> void
> dns_open ()
> {
> int option, i;
> res_init ();
> if (!_res.nscount)
> {
> fprintf (stderr, "No nameservers defined.\n");
> exit (-1);
> }
> ... other stuff ...
>
>
> So evidently "res_init();" sets its return or something or other
> in a sneaky (undocumented) global variable _res.nscount; if it's
> zero show you the error. Bad style. _res.nscount is a count, the
> test should be if(_res.nscount == 0) and no, that's not nit-picking.
>
> Let's look at this res_init() villain -- it's part of the resolver
> library, i.e. the C library. There's a man page for it.
>
> That manpage says nothing about success or failure of res_init().
> It does not say what res_init() might do.
>
> Well, consider li'l foo.c here:
>
> #include <sys/types.h>
> #include <netinet/in.h>
> #include <arpa/nameser.h>
> #include <resolv.h>
>
> main()
> {
> res_init();
> printf("%u\n", _res.nscount);
> return 0;
> }
>
> compiled and run on the same box that etherape complains about,
> it prints "1". Hmmmm.
>
> I don't know what's causing this error. My suspicion is that this
> gnome stuff, bonoboware, whatever, is compiled with threads and that
> it is relying on this static variable "_res", defined in libc, to
> be useful in more than one thread.
>
> Well, I went to dns.c, blocked out the error-checking for _res.nscounts,
> and the thing ran, but still didn't print the names of the hosts,
> just the IP numbers.
>
> It is not clear to me that the author is without blame. I'd suggest
> talking to him.
>
> The other errors at run-time would be sufficient to make me delete the
> package.
>
> Can I suggest 'ethereal' instead? This has been withdrawn from the
> ports for some reason. I'm currently sucking in the source code
> and will report back later if it works. I recall it used to.
>
> Sorry for my general nastiness but I've sunk about six hours in this,
> only to discover undocumented bloated Linuxware. You can tell I don't
> like gnome, Linux, or eyecandy. From what I can guess from the undocumented
> etherape program, one could get the same information from "netstat",
> or write a much smaller program to display the same information
> using ncurses(3). Using 33% of a pretty snappy CPU to display what
> it's displaying is simply inexcusable. Test and admin ware should
> not impinge on the operating system. Most of the stuff etherape
> reported to me looked like its own traffic to the Xserver. Tsk. Tsk.
>
> Please tell me exactly what you want from etherape, and I'll try to
> figure out another way to get it. I don't want to fix its bug with
> the resolver. The maintainer of the etherape port is
>
> Craig Barraclough <[EMAIL PROTECTED]>
>
> I think you might do well to ask him about this resolver problem,
> tell him it is still present in 4.0.
>
> If you *have* to have it, and it *has* to work, install the red-hat
> Linux emulation, and install a Linux binary of etherape (and all its
> gnome crap), and run it that way. THe preceding applies to other
> Linux programs.
>
> Dave
> _______________________________________________
> Openbsd-newbies mailing list
> [email protected]
> http://mailman.theapt.org/listinfo/openbsd-newbies
>
Hello,
I had the same problem a year or so ago, with etherape and the lack of dns
http://marc.theaimsgroup.com/?l=openbsd-misc&m=111465469331179&w=2
To get around it you can find a patch here for 0.91
http://www.networkpenetration.com/downloads.html
Basically it adds a -D switch so you can specify the DNS server.....
be warned though its a cpu hog and it fragged a machine of mine after
a few weeks of constant running.
Cheers
Ste
_______________________________________________
Openbsd-newbies mailing list
[email protected]
http://mailman.theapt.org/listinfo/openbsd-newbies
