Makefile.PL for Pg.so

Larry Rosenman writes:

I made the following patch, and it works for MY platform.

Peter,
Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

This is an interesting idea. I'd rather rip out MakeMaker completely, but
this might be a good start.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

* Peter Eisentraut <peter_e@gmx.net> [010814 16:58]:

Larry Rosenman writes:

I made the following patch, and it works for MY platform.

Peter,
Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

This is an interesting idea. I'd rather rip out MakeMaker completely, but
this might be a good start.

I'm not familiar enough with the config / autoconf stuff to do a
portable patch. Can you help in this area?

LER

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

* Peter Eisentraut <peter_e@gmx.net> [010825 18:14]:

Larry Rosenman writes:

Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

AFAICT, Pg.so does get the runpath set correctly. Are you saying it
doesn't work on your system or do you want to get rid of the
recompilation during installation?

It doesn't work on this systsm .

(you still have a login here).

LER

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

Larry Rosenman writes:

Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

AFAICT, Pg.so does get the runpath set correctly. Are you saying it
doesn't work on your system or do you want to get rid of the
recompilation during installation?

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

* Peter Eisentraut <peter_e@gmx.net> [010825 18:14]:

Larry Rosenman writes:

Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

AFAICT, Pg.so does get the runpath set correctly. Are you saying it
doesn't work on your system or do you want to get rid of the
recompilation during installation?

Oh, BTW, LD_RUN_PATH is *NOT* used to set the runpath on SVR5. That's
why I needed to pass -R to the build.

LER

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

On Saturday 25 August 2001 19:18, Peter Eisentraut wrote:

AFAICT, Pg.so does get the runpath set correctly. Are you saying it
doesn't work on your system or do you want to get rid of the
recompilation during installation?

Pg.so does not get the proper RPATH in a DESTDIR build environment.

Trying to fix this for the RPMS -- the RPATH contains the buildroot instead
of where the libs really are. Could cause security problems. Working on it,
but it's slow going for me at this busy time.
--
Lamar Owen
WGCR Internet Radio
1 Peter 4:11

Lamar Owen writes:

Pg.so does not get the proper RPATH in a DESTDIR build environment.

Trying to fix this for the RPMS -- the RPATH contains the buildroot instead
of where the libs really are. Could cause security problems. Working on it,
but it's slow going for me at this busy time.

Another fun feature of the DESTDIR build environment is that the
writability test of the target directory will most likely fail because it
doesn't exist at all.

I've been thinking how I'd like to fix this: We add an option to
configure that says to *not* install the Perl module into the standard
Perl tree, but instead somewhere under our own $prefix. That way people
that don't have root access can use this option and still install the
whole tree in one run. But then we'd remove that writability check and
people that have root access or failed to use that option will get a hard
failure. This would create a much more reliable and predictable build
environment.

Comments? And a good name for such an option?

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

Peter Eisentraut <peter_e@gmx.net> writes:

Another fun feature of the DESTDIR build environment is that the
writability test of the target directory will most likely fail because it
doesn't exist at all.

I've been thinking how I'd like to fix this: We add an option to
configure that says to *not* install the Perl module into the standard
Perl tree, but instead somewhere under our own $prefix. That way people
that don't have root access can use this option and still install the
whole tree in one run. But then we'd remove that writability check and
people that have root access or failed to use that option will get a hard
failure. This would create a much more reliable and predictable build
environment.

Why would we remove the writability check? Perhaps it needs to be
extended to recognize the case of target-dir-doesn't-exist-but-can-be-
created, but I don't see why a hard failure is better.

regards, tom lane

* Tom Lane <tgl@sss.pgh.pa.us> [010826 11:11]:

Peter Eisentraut <peter_e@gmx.net> writes:

Another fun feature of the DESTDIR build environment is that the
writability test of the target directory will most likely fail because it
doesn't exist at all.

I've been thinking how I'd like to fix this: We add an option to
configure that says to *not* install the Perl module into the standard
Perl tree, but instead somewhere under our own $prefix. That way people
that don't have root access can use this option and still install the
whole tree in one run. But then we'd remove that writability check and
people that have root access or failed to use that option will get a hard
failure. This would create a much more reliable and predictable build
environment.

Why would we remove the writability check? Perhaps it needs to be
extended to recognize the case of target-dir-doesn't-exist-but-can-be-
created, but I don't see why a hard failure is better.

I tend to agree with Tom here. The original problem I was seeing is
*NOT* related to DESTDIR (I don't believe).

CC=cc CXX=CC ./configure --prefix=/usr/local/pgsql --enable-syslog \
--with-CXX --with-perl --enable-multibyte --enable-cassert \
--with-includes=/usr/local/include --with-libs=/usr/local/lib \
--enable-debug \
--with-tcl --with-tclconfig=/usr/local/lib \
--with-tkconfig=/usr/local/lib --enable-locale --with-python

The above is my configure input.

regards, tom lane

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

Lamar Owen <lamar.owen@wgcr.org> writes:

On Saturday 25 August 2001 19:18, Peter Eisentraut wrote:

AFAICT, Pg.so does get the runpath set correctly. Are you saying it
doesn't work on your system or do you want to get rid of the
recompilation during installation?

Pg.so does not get the proper RPATH in a DESTDIR build environment.

Rather: Perl decides it wants to specify a LD_RUN_PATH in its makefile.
This will automatically make use of -R. It's fixed in the RPMs
available at http://people.redhat.com/teg/pg/

--
Trond Eivind Glomsr�d
Red Hat, Inc.

Trond Eivind Glomsr�d writes:

Pg.so does not get the proper RPATH in a DESTDIR build environment.

Rather: Perl decides it wants to specify a LD_RUN_PATH in its makefile.
This will automatically make use of -R. It's fixed in the RPMs
available at http://people.redhat.com/teg/pg/

AFAICT, these RPMs still contain the same problem that the runpath will
point the the temporary installation directory, not the final location.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

Peter Eisentraut <peter_e@gmx.net> writes:

Trond Eivind Glomsr�d writes:

Pg.so does not get the proper RPATH in a DESTDIR build environment.

Rather: Perl decides it wants to specify a LD_RUN_PATH in its makefile.
This will automatically make use of -R. It's fixed in the RPMs
available at http://people.redhat.com/teg/pg/

AFAICT, these RPMs still contain the same problem that the runpath will
point the the temporary installation directory, not the final location.

Doh... these were just temporary rpms, I've put the final -2 rpms
there now (built later the same day).

Basically, I fixed it by removing references to LD_RUN_PATH after perl
had generated it.
--
Trond Eivind Glomsr�d
Red Hat, Inc.

* Peter Eisentraut <peter_e@gmx.net> [010826 17:42]:

I've checked in a fix for the runpath, DESTDIR, and VPATH problems. The
first needs to be tested on a variety of platforms because I had to guess
the osname configuration values and the linker/compiler that Perl uses.

I ass/u/me that you tested it on lerami?

LER

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

I've checked in a fix for the runpath, DESTDIR, and VPATH problems. The
first needs to be tested on a variety of platforms because I had to guess
the osname configuration values and the linker/compiler that Perl uses.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

Can someone comment on this?

I made the following patch, and it works for MY platform.

Peter,
Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

Index: Makefile.PL
===================================================================
RCS file: /home/projects/pgsql/cvsroot/pgsql/src/interfaces/perl5/Makefile.PL,v
retrieving revision 1.17
diff -c -r1.17 Makefile.PL
*** Makefile.PL	2001/03/06 22:07:09	1.17
--- Makefile.PL	2001/08/13 04:12:28
***************
*** 64,66 ****
--- 64,77 ----
];

}
+ sub MY::dynamic_lib {
+       package MY;
+       my $inherited= shift->SUPER::dynamic_lib(@_);
+       if (! -d $ENV{POSTGRES_LIB} ) {
+        my $cwd = `pwd`;
+       chop $cwd;
+       $ENV{POSTGRES_LIB} = "$cwd/../libpq";
+       }
+       $inherited=~ s@OTHERLDFLAGS =@OTHERLDFLAGS =-R$ENV{POSTGRES_LIB}@;
+       $inherited;
+ }

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo@postgresql.org)

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

I think peter_e dealt with this....

LER

* Bruce Momjian <pgman@candle.pha.pa.us> [010907 13:25]:

Can someone comment on this?

I made the following patch, and it works for MY platform.

Peter,
Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

Index: Makefile.PL
===================================================================
RCS file: /home/projects/pgsql/cvsroot/pgsql/src/interfaces/perl5/Makefile.PL,v
retrieving revision 1.17
diff -c -r1.17 Makefile.PL
*** Makefile.PL	2001/03/06 22:07:09	1.17
--- Makefile.PL	2001/08/13 04:12:28
***************
*** 64,66 ****
--- 64,77 ----
];

}
+ sub MY::dynamic_lib {
+       package MY;
+       my $inherited= shift->SUPER::dynamic_lib(@_);
+       if (! -d $ENV{POSTGRES_LIB} ) {
+        my $cwd = `pwd`;
+       chop $cwd;
+       $ENV{POSTGRES_LIB} = "$cwd/../libpq";
+       }
+       $inherited=~ s@OTHERLDFLAGS =@OTHERLDFLAGS =-R$ENV{POSTGRES_LIB}@;
+       $inherited;
+ }

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo@postgresql.org)

-- 
Bruce Momjian                        |  http://candle.pha.pa.us
pgman@candle.pha.pa.us               |  (610) 853-3000
+  If your life is a hard drive,     |  830 Blythe Avenue
+  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

Yes, thanks. Cleaning out my mailbox, as usualy before a beta.

I think peter_e dealt with this....

LER

* Bruce Momjian <pgman@candle.pha.pa.us> [010907 13:25]:

Can someone comment on this?

I made the following patch, and it works for MY platform.

Peter,
Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

Index: Makefile.PL
===================================================================
RCS file: /home/projects/pgsql/cvsroot/pgsql/src/interfaces/perl5/Makefile.PL,v
retrieving revision 1.17
diff -c -r1.17 Makefile.PL
*** Makefile.PL	2001/03/06 22:07:09	1.17
--- Makefile.PL	2001/08/13 04:12:28
***************
*** 64,66 ****
--- 64,77 ----
];

}
+ sub MY::dynamic_lib {
+       package MY;
+       my $inherited= shift->SUPER::dynamic_lib(@_);
+       if (! -d $ENV{POSTGRES_LIB} ) {
+        my $cwd = `pwd`;
+       chop $cwd;
+       $ENV{POSTGRES_LIB} = "$cwd/../libpq";
+       }
+       $inherited=~ s@OTHERLDFLAGS =@OTHERLDFLAGS =-R$ENV{POSTGRES_LIB}@;
+       $inherited;
+ }

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo@postgresql.org)

-- 
Bruce Momjian                        |  http://candle.pha.pa.us
pgman@candle.pha.pa.us               |  (610) 853-3000
+  If your life is a hard drive,     |  830 Blythe Avenue
+  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

Bruce Momjian writes:

Can someone comment on this?

It's done.

I made the following patch, and it works for MY platform.

Peter,
Can we do something similar for the distribution to set the
RUNPATH for Pg.so?

Index: Makefile.PL
===================================================================
RCS file: /home/projects/pgsql/cvsroot/pgsql/src/interfaces/perl5/Makefile.PL,v
retrieving revision 1.17
diff -c -r1.17 Makefile.PL
*** Makefile.PL	2001/03/06 22:07:09	1.17
--- Makefile.PL	2001/08/13 04:12:28
***************
*** 64,66 ****
--- 64,77 ----
];

}
+ sub MY::dynamic_lib {
+       package MY;
+       my $inherited= shift->SUPER::dynamic_lib(@_);
+       if (! -d $ENV{POSTGRES_LIB} ) {
+        my $cwd = `pwd`;
+       chop $cwd;
+       $ENV{POSTGRES_LIB} = "$cwd/../libpq";
+       }
+       $inherited=~ s@OTHERLDFLAGS =@OTHERLDFLAGS =-R$ENV{POSTGRES_LIB}@;
+       $inherited;
+ }

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo@postgresql.org)

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter