Mailing List Archive

Compiles and tests runs ok on HP-UX 9.01 (HP700 workstation)

When installing I had the following Configure related problem:

The hint file suggests d_bsdpgrp='define', but
the configure test indicated that setpgrp(void) is the correct form

The problem seems to rise from the fact HP-UX (9.01) has the
functions setpgrp2 and getpgrp2 which will be used by the actual
perl code, but Configure does its testing by using plain setpgrp.

Eero

--
Eero.Pajarre@vlsi.fi VLSI Solution OY
Tel:+358 31 3165270 Kanslerinkatu 6
Fax:+358 31 3165220 FIN-33720 Tampere, Finland

lwall> Packrats, don't worry too much about readvertising this widely.
lwall> Nowadays we're on a T1 here, so our bandwidth is okay.

5.002beta1 in CPAN, over and out.

++jhi;

> The hint file suggests d_bsdpgrp='define', but
> the configure test indicated that setpgrp(void) is the correct form
>
> The problem seems to rise from the fact HP-UX (9.01) has the
> functions setpgrp2 and getpgrp2 which will be used by the actual
> perl code, but Configure does its testing by using plain setpgrp.

HP-UX is indeed strange. If you use 'define', it will still fail, but
it better follows the Principle of Least Astonishment.

The workaround is to use POSIX::setsid.

Jeff

Compiled an passed all tests on an HP 712, HP/UX 9.05.

Randy
--
^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^
Randy J. Ray -- U S WEST Technologies IAD/CSS/DPDS Phone: (303)595-2869
Denver, CO rjray@lookout.ecte.uswc.uswest.com

I don't suffer from insanity. I enjoy every minute of it.

Hello,

Linux 1.3.43, ELF, libc 5.2.14
Perl5.002beta1, Configure -des

Tests 100% succesful.

Congratulations!

--
Bart.

On Tue, 21 Nov 1995 10:37:03 PST, Larry Wall wrote:
>If you're adventurous, have a look at
>
> ftp://ftp.sems.com/pub/outgoing/perl5.0/perl5.002beta1.tar.gz
>
>Many thanks to Andy for doing the integration.
>

Passes with flying colors on both Linux 1.2.13 and SunOS 4.1.3.

Kudos to Larry, Andy, Tom and rest of the porters for the
fine job.

- Sarathy.
gsar@engin.umich.edu

P.S: While looking at the docs, this patent inaccuracy caught my
eye. Since I don't have the time to look at the code this
very minute, I tried fixing the docs instead :-)
------------------------------8<----------------------------
*** perlre.pod.dist Wed Nov 15 21:35:31 1995
--- perlre.pod Tue Nov 21 23:05:35 1995
***************
*** 75,82 ****

(If a curly bracket occurs in any other context, it is treated
as a regular character.) The "*" modifier is equivalent to C<{0,}>, the "+"
! modifier to C<{1,}>, and the "?" modifier to C<{0,1}>. There is no limit to the
! size of n or m, but large numbers will chew up more memory.

By default, a quantified subpattern is "greedy", that is, it will match as
many times as possible without causing the rest pattern not to match. The
--- 75,82 ----

(If a curly bracket occurs in any other context, it is treated
as a regular character.) The "*" modifier is equivalent to C<{0,}>, the "+"
! modifier to C<{1,}>, and the "?" modifier to C<{0,1}>. n and m are limited
! to integral values less than 65536.

By default, a quantified subpattern is "greedy", that is, it will match as
many times as possible without causing the rest pattern not to match. The

Just another couple of data points for 5.002 beta 1. It builds out of the
box and passes all tests on both flavor Harris Night Hawk systems:

CX/UX amber 7.1 AMBER.101095 m88k

Power_UNIX george 2.2 Power_UNIX Motorola ppc604
--
Tom.Horsley@mail.hcsc.com
Home: 511 Kingbird Circle Delray Beach FL 33444
Work: Harris Computers, 2101 W. Cypress Creek Rd. Ft. Lauderdale FL 33309
Support Project Vote Smart! They need your support in non-election years too!
(email pvs@neu.edu, 1-800-622-SMART, http://www.vote-smart.org)

thanks for the fix. i haven't gotten to perlre doc fixen yet.

--tom

> Message ID: <9511211837.AA12030@scalpel.netlabs.com>
> From: Larry Wall <lwall@scalpel.netlabs.com>
> Date: Tue Nov 21 11:37:03 1995
> Subject: 5.002 beta 1
> ------------------------------------------------------------
>
> If you're adventurous, have a look at
>
> ftp://ftp.sems.com/pub/outgoing/perl5.0/perl5.002beta1.tar.gz
>
> Many thanks to Andy for doing the integration.

Perl Version: perl5.002beta1 + patch-perl5.002beta1-MM5.09
System: SCO ODT 3.0 (or SCO Unix 3.2v4.2)

Ran into only one problem: ./ext/Socket/Socket.xs defines "<sys/un.h>" and
contains two routines based on <sys/un.h> which SCO Unix does not support.

./../../miniperl -I../../lib -I../../lib ../../lib/ExtUtils/xsubpp -typemap
../../lib/ExtUtils/typemap Socket.xs >Socket.tc && mv Socket.tc Socket.c
cc -c -W0 -UM_XENIX -DDEBUGGING -g -DVERSION=\"1.3\" -I../.. Socket.c
Socket.c
Socket.c(18) : fatal error C1024: cannot open include file 'sys/un.h'

The attached "patch" removes the offending code from "Socket.xs" and
the compilation succeeds without any further problems, and all tests pass.
This is obviously not the ideal solution. I'm afraid I don't know anything
about *.xs code, so I didn't attempt to make a more generic patch using
#ifndef's, etc. I'll leave that to the experts. :)

Thanks for an excellent job once again!!

...Bill
--
Bill Aten <bill@netagw.com> Member of UVV
==============================================================================
"Nice boy, but about as sharp as a sack of wet mice." -- Foghorn Leghorn
==============================================================================

Index: ./perl5.002beta1/ext/Socket/Socket.xs
*** ./perl5.002beta1/ext/Socket/Socket.xs~ Sat Nov 18 13:36:57 1995
--- ./perl5.002beta1/ext/Socket/Socket.xs Thu Nov 23 14:56:39 1995
***************
*** 7,13
# include <sys/types.h>
# endif
#include <sys/socket.h>
- #include <sys/un.h>
# ifdef I_NETINET_IN
# include <netinet/in.h>
# endif

--- 7,12 -----
# include <sys/types.h>
# endif
#include <sys/socket.h>
# ifdef I_NETINET_IN
# include <netinet/in.h>
# endif
***************
*** 625,668
addr_str = inet_ntoa(addr);

ST(0) = sv_2mortal(newSVpv(addr_str, strlen(addr_str)));
- }
-
- void
- pack_sockaddr_un(pathname)
- char * pathname
- CODE:
- {
- struct sockaddr_un sun_ad; /* fear using sun */
- Zero( &sun_ad, sizeof sun_ad, char );
- sun_ad.sun_family = AF_UNIX;
- Copy( pathname, sun_ad.sun_path, sizeof sun_ad.sun_path, char );
- ST(0) = sv_2mortal(newSVpv((char *)&sun_ad, sizeof sun_ad));
- }
-
- void
- unpack_sockaddr_un(sun_sv)
- SV * sun_sv
- PPCODE:
- {
- STRLEN sockaddrlen;
- struct sockaddr_un addr;
- char * sun_ad = SvPV(sun_sv,sockaddrlen);
-
- if (sockaddrlen != sizeof(addr)) {
- croak("Bad arg length for %s, length is %d, should be %d",
- "Socket::unpack_sockaddr_un",
- sockaddrlen, sizeof(addr));
- }
-
- Copy( sun_ad, &addr, sizeof addr, char );
-
- if ( addr.sun_family != AF_UNIX ) {
- croak("Bad address family for %s, got %d, should be %d",
- "Socket::unpack_sockaddr_un",
- addr.sun_family,
- AF_UNIX);
- }
- ST(0) = sv_2mortal(newSVpv(addr.sun_path, strlen(addr.sun_path)));
}

void

--- 624,629 -----
addr_str = inet_ntoa(addr);

ST(0) = sv_2mortal(newSVpv(addr_str, strlen(addr_str)));
}

void
==============================================================================

> Message ID: <9511240807.AA07602@aachen.ericsson.se>
> From: Tom Christiansen <teacher@aachen.eed.ericsson.se>
> Date: Fri Nov 24 01:07:00 1995
> Subject: Re: 5.002 beta 1
> ------------------------------------------------------------
>
> >Ran into only one problem: ./ext/Socket/Socket.xs defines "<sys/un.h>" and
> >contains two routines based on <sys/un.h> which SCO Unix does not support.
>
> >- pack_sockaddr_un(pathname)
> >- unpack_sockaddr_un(sun_sv)
>
> Don't you dare.
>
> --tom

Tom,

If I understand your comment correctly, then you didn't read my message
closely. I was not advocating removing this support from the package for
everybody. I said I didn't know *.xs well enough to do the proper "#ifdef"
or "#ifndef" patch which would only remove the code for SCO 3.2v4.2 systems
(I don't know if SCO OSE 5.0 supports <sys/un.h> or not at the moment). I
included the patch that I used to get things working for SCO so whoever
attempted to accomplish the *proper* fix would know exactly which section of
code I was referring to in my message.

Sorry for any confusion...

...Bill
--
Bill Aten <bill@netagw.com> Member of UVV
==============================================================================
No good deed goes unpunished. -- Clare Boothe Luce

>Ran into only one problem: ./ext/Socket/Socket.xs defines "<sys/un.h>" and
>contains two routines based on <sys/un.h> which SCO Unix does not support.

>- pack_sockaddr_un(pathname)
>- unpack_sockaddr_un(sun_sv)

Don't you dare.

--tom

I haven't seen this trivial problem reported yet...

$ uname -a
FreeBSD grommet.pipex.net 2.0.5-RELEASE FreeBSD 2.0.5-RELEASE #0: Tue Nov 7 11:29:45 GMT 1995 root@grommet.pipex.net:/usr/src/sys/compile/GROMMET i386
$ sh -n Configure
Configure: 5419: Syntax error: ";" unexpected

Having fixed this (patch appended, for what it's worth), it built
smoothly and passed all tests.

Tim.

--- Configure-dist Mon Nov 20 15:00:33 1995
+++ Configure Fri Nov 24 14:37:17 1995
@@ -5433,7 +5433,7 @@
exit(0); /* Bye bye, thank you for playing! */
}
EOCP
- if $cc $ccflags $ldflags try.c -o try >/dev/null; 2>&1; then
+ if $cc $ccflags $ldflags try.c -o try >/dev/null 2>&1; then
echo "./try >try.out 2>try.ret 3>try.err || exit 4" >mtry
chmod +x mtry
./mtry >/dev/null 2>&1

On Fri, 24 Nov 1995, Tim Goodwin wrote:

> I haven't seen this trivial problem reported yet...
>
> $ uname -a
> FreeBSD grommet.pipex.net 2.0.5-RELEASE FreeBSD 2.0.5-RELEASE #0: Tue Nov 7 11:29:45 GMT 1995 root@grommet.pipex.net:/usr/src/sys/compile/GROMMET i386
> $ sh -n Configure
> Configure: 5419: Syntax error: ";" unexpected
>
> Having fixed this (patch appended, for what it's worth), it built
> smoothly and passed all tests.

Excellent detective work. Thanks. Many pairs of eyes have missed that
extra comma. (The wrong line number in Configure threw us off.)

Andy Dougherty doughera@lafcol.lafayette.edu
Dept. of Physics
Lafayette College, Easton PA 18042