Re: Seeking advice on oddities in VIS installation process

From: Roderick Bloem (rbloem@yahoo.com)
Date: Fri Nov 24 2000 - 10:16:29 MST

  • Next message: Kathi Fisler: "Re: Seeking advice on oddities in VIS installation process"

    Hi Kathi,

    Thanks for your feedback.

    We have not seen this problem before. I thought at first that the
    problem might be Solaris 2.8, it is apparently a rather big overhaul
    with 64 bit addressing for the first time.

    However, I tried to install vis-1.3 on our 2.8 machine, and it worked
    fine, both using cc and using gcc (though there are rather many
    warnings).

    We use gcc 2.8.1 and gmake 3.77.

    Do you have access to an older gcc version? Maybe we can see if we can
    upgrade ours to 2.95.*. The spurious files sound like something may
    well be wrong in the make process.

    Could you perhaps send us your makefile?

    Roderick.

    Kathi Fisler wrote:
    >
    > I am trying to install VIS 1.3 on an UltraSparc running Solaris 5.8.
    > I'm using gmake version 3.75 and gcc version 2.95.2. VIS compiles
    > easily, but about half of the tests from "make check" fail (all the
    > GLU 1.3 checks pass).
    >
    > I've noticed two differences between this installation attempt and the
    > many other times that I've (successfully) installed VIS. I'm
    > wondering whether anyone has insight into what causes these two
    > problems, in hope that they affect the failed checks.
    >
    > 1. Both the GLU and VIS installations create a slew of files in the
    > prefix directory specified in the Makefile. For example, for vmVers,
    > the following files appear:
    >
    > vmVers.c.addressof, vmVers.c.bp, vmVers.c.combine, vmVers.c.cse,
    > vmVers.c.cse2, vmVers.c.dbr, vmVers.c.flow, vmVers.c.flow2,
    > vmVers.c.gcse, vmVers.c.greg, vmVers.c.jump, vmVers.c.jump2,
    > vmVers.c.loop, vmVers.c.lreg, vmVers.c.regmove, vmVers.c.rtl,
    > vmVers.c.sched, and vmVers.c.sched2
    >
    > I get similar listings for most (if not all) of the .c files in each
    > distribution. What generates these files and is this a problem?
    >
    > 2. I get the following warning on every file compilation:
    >
    > cc1: warning: unrecognised gcc debugging option: i
    > cc1: warning: unrecognised gcc debugging option: n
    >
    > I don't find these options specified anywhere in the Makefile. Any
    > idea where they might be coming from?
    >
    > (Note: We don't have a native cc installed, so gcc is my only option.
    > I did run configure with --enable-gcc for the GLU and VIS
    > installations.)
    >
    > Thanks,
    > Kathi Fisler
    >
    > -----------------------------------------
    > Kathi Fisler, Assistant Professor
    > Department of Computer Science
    > Worcester Polytechnic Institute
    > http://www.cs.wpi.edu/~kfisler/

    _________________________________________________________
    Do You Yahoo!?
    Get your free @yahoo.com address at http://mail.yahoo.com



    This archive was generated by hypermail 2b29 : Fri Nov 24 2000 - 16:01:15 MST