Subject: Re: installing savant prerelease
From: Dale E Martin (dmartin@cliftonlabs.com)
Date: Mon Oct 22 2001 - 05:21:21 MDT
> Dale Martin asked me to submit bug reports; however, the
> bugzilla form has only the prerelease version as a possible
> version, and I am using 1.02. So, I decided to try installing
> the prerelease version, however I'm having problems.
>
> First, there is not INSTALL-FOR-SIMULATION file and the
> installation doesn't seem to go similarly as with 1.02.
> More specifically, "make install" in Tyvis doesn't seem to do
> anything.
Hi, you're right - we switched to an autotools build. I posted
instructions on this list back when we started the series. I'll be
updating the documentation very soon as things are starting to settle down
in the build system - the last issues are now being addressed. Here were
the directions posted back then:
http://www.cliftonlabs.com/savant/list-archives/savant-users/0020.html
To address the shared library issue, the next prerelease is going to
require libtool to be installed. I still need to play with a distribution
some to see what this actually means to the end user - perhaps it won't
even matter. We've added very rudimentary support for foreign
architectures and processes, which definitely needs libtool headers laying
around. I'm not sure if "make dist" will install those or not.
> Second, compiling Savant stops to these error messages:
>
> g++ -DPACKAGE=\"savant\" -DVERSION=\"2.0-pre010829\" -DYYTEXT_POINTER=1
> -DSTDC_HEADERS=1 -DHAVE_UNISTD_H=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1
> -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1
> -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DSIZEOF_LONG=4
> -DRETSIGTYPE=void -DHAVE_MKDIR=1 -DHAVE_STRDUP=1 -I. -I.
> -I/usr/src/pccts/h -I../../src -I../../src/aire
> -I../../src/aire/iir/visitor -I../../src/aire/iir/IIR
> -I../../src/aire/iir/IIRBase -I../../src/aire/iir/IIRScram
> -I../../src/visitors/module -I../../src/util
> -I../../src/analyzer -DDEVELOPER_ASSERTIONS -DGENERATE_CC_REFERENCES
> -DSIZEOF_LONG=4 -DHAS_BOOLEAN -g -Wall -c vhdl.cc
> vhdl.cc:632: macro `zzsetmatch' used with just one arg
> vhdl.cc:2200: macro `zzsetmatch' used with just one arg
> vhdl.cc:2405: macro `zzsetmatch' used with just one arg
> ...
>
> This is IT Linux 2000 (Redhat 6.2 clone), egcs-2.91.66,
> x86.
>
> I have PCCTS 1.33mr23.
Yes, PCCTS 1.33mr21-25 had a bug in semantic predicates that caused the
problem you're seeing. Grab the latest (mr31) and you should be OK -
here's the URL:
http://www.polhode.com/pccts133mr.html
I should also say that this code has never been tested with that version of
g++. We're using STL, namespaces, and some of the newer features of C++ so
it's possible that you'll see some problems. I would really like to hear
how it goes either way.
Thanks,
Dale
-- Dale E. Martin, Clifton Labs, Inc. Senior Computer Engineer dmartin@cliftonlabs.com http://www.cliftonlabs.com pgp key available
This archive was generated by hypermail 2b25 : Mon Mar 18 2002 - 13:00:02 MST