Re: bison-1.29 build failures along the line

[H . Merijn Brand]

On Thu 11 Oct 2001 12:27, Akim Demaille <address@hidden> wrote:
> >>>>> "H" == H Merijn Brand <H.Merijn> writes:
> 
> * HP
> 
> What the heck is this message:
> 
> 13. calc.at:383: testing Calculator ...
> calc.at:383: bison calc.y -o calc.c 
> calc.at:383: $CC $CFLAGS $CPPFLAGS calc.c -o calc
> --- /dev/null   Wed Oct 10 17:47:17 2001
> +++ at-stderr   Wed Oct 10 17:47:18 2001
> @@ -0,0 +1 @@
> +/usr/ccs/bin/ld: (Warning) At least one PA 2.0 object file (calc.o) was 
> detecte
> d. The linked output may not run on a PA 1.x system.

The default build on HP-UX systems compiles for the native architecture. For
modern HP9000 systems with hppa processors, that'll be hppa-2.0w

Objects built for hppa-2.0 will not work on older processor family types like
hppa-1.1 and hppa-1.0, even when running the same OS (e.g. HP-UX 11.00)

My Policy.sh has 'LDFLAGS="-Wl,+vnocompatwarnings"', which suppresses this
message, but somewhere in the make process, this LDFLAGS isn't applied ;(

The message is however totally harmless for native builds :)

> I consider the test suite is fine here, it is just your system which
> is bogus :(  I guess it means the test suite should check the warnings
> only if running gcc :(

I guest this is an OK to 'make install' ?

> * AIX
> 
> 13. calc.at:383: testing Calculator ...
> calc.at:383: bison calc.y -o calc.c 
> calc.at:383: $CC $CFLAGS $CPPFLAGS calc.c -o calc
> --- /dev/null   Wed Oct 10 20:01:38 2001
> +++ at-stderr   Wed Oct 10 20:02:02 2001
> @@ -0,0 +1 @@
> +1506-507 (W) No licenses available. Contact your program supplier to add 
> additi
> onal users.  Compilation will proceed shortly.

This is something completely beyond me :)
I've upgraded vac-5.0.2.0 which ran fine for one user limit to vac-5.0.2.1 and
as of that moment I get warnings, *but* is runs fine (except for that annoying
message) for an unlimited number of users. I've contacted our supplyer about
that, but they said I should contact IBM, but IBM does do absolutely nothing
for free, and since the message doesn't harm the results, I'm (we are) not
willing to pay for it.

> 13. calc.at:383: FAILED near `calc.at:383'
> 
> OK, same problem, can't trust the output of your compiler.  Thanks,
> Bison is fine :)

Can we ever trust IBM's output ever? :)

-- 
H.Merijn Brand    Amsterdam Perl Mongers (http://www.amsterdam.pm.org/)
using perl-5.6.1, 5.7.1 & 629 on HP-UX 10.20 & 11.00, AIX 4.2, AIX 4.3,
     WinNT 4, Win2K pro & WinCE 2.11 often with Tk800.022 &/| DBD-Unify
ftp://ftp.funet.fi/pub/languages/perl/CPAN/authors/id/H/HM/HMBRAND/