[ARGUS] build with bison OR yacc

[Carter Bullard]

Well, the only reason we shifted to bison so long ago was
because yacc sets a large number of globals, and when we
linked with libpcap, which also used yacc at the time,
we would get multiple defines of the same global values.

Bison fixed this a long time ago, by allowing you to set
a global variable prefix, but yacc took its time,
and some vendors, like SGI, took even longer to get yacc
to where you could have multiple compilers linked together.

I'll add Mike's ./configure mods, we'll put yacc up front,
and we'll see if anyone complains!!!

Carter 




> From: eric <eric-list-argus at catastrophe.net>
> Organization: Catastrophe.Net <http://www.catastrophe.net/>
> Date: Thu, 30 Sep 2004 12:25:00 -0500
> To: <slif at bellsouth.net>
> Cc: Carter Bullard <carter at qosient.com>, Peter Van Epp <vanepp at sfu.ca>, Argus
> <argus-info at lists.andrew.cmu.edu>
> Subject: Re: Re: [ARGUS] build with bison OR yacc
> 
> On Thu, 2004-09-30 at 13:16:35 -0400, slif at bellsouth.net proclaimed...
> 
>> My point of offering the patch is building on FreeBSD 5.2.1
>> appears to work well without bison.
>> 
>> Hey, the configure variable is named V_YACC, that was a big clue...
>> 
>> Given what you've just described,
>> the configure test should check for -p support
>> for yacc, and fall back to expecting bison, throwing
>> an error only when neither of those methods work.
> 
> I'd love to see yacc become the standard; bison is about the only
> package that I install on my sensors, and if I could just save that
> extra 412kbytes of disk space! :)
>