Peter Sinnott wrote:
On Thu, Jun 02, 2005 at 11:22:45AM +0100, Nick Glencross wrote:
Peter Sinnott wrote:
bash-2.05$ uname -a
HP-UX gnbil2dv B.11.00 A 9000/800 1657309373 two-user license
bash-2.05$ /cm/tools/bin/perl -V
Summary of my perl5 (revision 5.0 version 8 subversion 0) configuration:
Platform:
osname=hpux, osvers=11.00, archname=PA-RISC2.0
uname='hp-ux cmhp01 b.11.00 a 9000800 131901507 two-user license '
I guess that just doing a s|/|_|g would be a good start.
When I get access, I'll also see if the alignment test hangs as one of
the comments suggests, or whether I get a meaningful value returned.
From cvs :
Determining your minimum pointer alignment..............for hpux: 4 bytes.
With hpux hard code commented out :
Determining your minimum pointer alignment........................ 4 bytes.
The comment in the hardcode says it is for HP-UX 10.20 so maybe 11 never
hung. Same answer anyway so no harm done(probably)
Great! That saves me doing tomorrow's experiment. Have you done a build
to see if you get similar behaviour to me?
Alas my attempts to build fail at the link stage. gcc 2.95 fails
as stuff isn't compiled correctly with position independent code and
nothing I try with fpic/fPIC/shared will please it. gcc 3.3 fails
saying there are references to variables that do not exist and the
variable looks like some scary internal hp(not parrot) variable as it
starts with a _. I may try building again tomorrow but probably will
play about with the solaris failures today.
Is this in the dynclasses directory or the nci stuff?
If so, in the top-level makefile, find the 'all :' section and remove
the dependency on $(LIBNCI_TEST_SO) by just removing the line.
Then in dynclasses/Makefile, just replace the whole file with 'all:'.
That should allow the parrot binary to be built,
Nick
