checking for Linux LFS... ./configure: test: =: unary operator expected (was "Re: argus-2.0.5.beta.1")

[Carter Bullard]

Hey Dave,
   Thanks!  When I looked a few months ago, AC_SYS_LARGEFILE
did not exist, so yes, I definitely would rather use that.
The best thing to do right now, is just turn it off, and
not generate files that are tooooo big.

   I've run into an interesting problem with FlowScan, not
a biggy, just a head scratcher.  Seems that when using one
argus distribution, "struct block" ends up being defined
mutliple times, and in another its ok.  Argus defines a
"struct block" for its internal record filter compiler,
but I can't find any  reference to a "struct block" in
any of the FlowScan stuff.

Any ideas?  (I just renamed the struct to "ablock" in 
offending distribution, so not a major thing, just
interesting).

Carter

Carter Bullard
QoSient, LLC
300 E. 56th Street, Suite 18K
New York, New York  10022

carter at qosient.com
Phone +1 212 588-9133
Fax   +1 212 588-9134
http://qosient.com

> -----Original Message-----
> From: owner-argus-info at lists.andrew.cmu.edu 
> [mailto:owner-argus-info at lists.andrew.cmu.edu] On Behalf Of 
> Dave Plonka
> Sent: Monday, January 28, 2002 6:05 PM
> To: argus-info at lists.andrew.cmu.edu
> Subject: checking for Linux LFS... ./configure: test: =: 
> unary operator expected (was "Re: argus-2.0.5.beta.1")
> 
> 
> On Wed, Jan 16, 2002 at 09:48:28AM -0500, Carter Bullard wrote:
> > Gentle people,
> >    I've just uploaded argus-2.0.5.beta.1 to the dev directory. This 
> > will be our testing distribution for the next release.
> 
> Carter,
> 
> I noticed a problem with the "configure" in 
> argus-2.0.5.beta.1 under Linux.  Specifically, on a Debian 
> GNU/Linux machine:
> 
>    $ ./configure       
>    creating cache ./config.cache
>    checking host system type... i686-pc-linux-gnu
>    ...
>    checking for Linux LFS... ./configure: test: =: unary 
> operator expected
>    ./configure: test: -lt: unary operator expected
>    ./configure: test: -lt: unary operator expected
> 
> I poked around in "configure.in" and see the problem is here:
> 
>    AC_CACHE_VAL(ac_cv_linux_vers,
>             ac_cv_linux_vers=`uname -r 2>&1 | \
>             sed -n -e '$s/.* //' -e '$s/\-.*//p'`)
> 
> Apparently that's not a portable way (amongst Linux versions) 
> to determine the version numbers because "uname -r" doesn't 
> always contain a "-".  E.g. on my machine:
> 
>    $ uname -r
>    2.2.18pre21
>    $
> 
> (Linux 2.2.18pre21 is the Linux kernel used in the current 
> stable Debian 2.2 release.)
> 
> I see in the latest version of autoconf that there is this 
> (new?) macro:
> 
>    AC_SYS_LARGEFILE
> 
> Perhaps that could be used instead, and avoid argus from 
> having to try to determine the Linux version numbers?  Sorry, 
> I don't have a patch though... I played with it for couple 
> minutes but had other problems with getting "autoconf-2.52" to work.
> 
> If you do end up hand-crafting the test, the AC_SYS_LARGEFILE 
> macro might be useful for comparison.  It is defined in 
> "autoconf-2.52/acspecific.m4".
> 
> Dave
> 
> -- 
> plonka at doit.wisc.edu  http://net.doit.wisc.edu/~plonka  
> ARS:N9HZF  Madison, WI
> 
>