Обсуждение: urgent: pgsql 6.3 FreeBSD port update trouble, --with-tcl doesn't work

Hi !

Please help, the FreeBSD port update to version 6.3 drives me to
despair ;-)

--with-tcl configure switch doesn't work to build libpgtcl

I only get it to work if I set in the make environment

    USE_TCL=true

configure --with-tcl doesn't seem to set USE_TCL=true in
src/Makefile.global, so
    src/interfaces/Makefile,
which includes
    src/Makefile.global
fails, since USE_TCL isn't set to true. Workaround was to explicitely
set USE_TCL=true when starting the make.

Another problem: on FreeBSD-current (and I think this is true on
FreeBSD-2.2.x as well)

The make install fails when running the initdb script.

Adding template1 database to pg_database...
Running: postgres -boot -C -F -D/usr/local/pgsql/data -Q template1 < /tmp/create
.13645

vacuuming template1
creating public pg_user view
loading pg_description
*** Error code 1

The Postgresql 6.2 version needed a patch so that this works
--- bin/initdb/initdb.sh.orig   Wed Apr  2 07:09:13 1997
+++ bin/initdb/initdb.sh        Wed Apr  2 07:09:33 1997
@@ -344,6 +344,5 @@

 if [ $debug -eq 0 ]; then
     echo "vacuuming template1"
-    echo "vacuum" | postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
-        grep -v "^DEBUG:"
+    echo "vacuum" | postgres -o /dev/null -F -Q -D$PGDATA template1 > /dev/null

 fi

I enhanced this patch and now use
    "-o /dev/null"
and avoid
    "2>&1"

several lines below where similar commands can be found, but this
doesn't solve the problem.

--- bin/initdb/initdb.sh.orig   Sun Mar 22 12:47:51 1998
+++ bin/initdb/initdb.sh        Sun Mar 22 13:06:36 1998
@@ -348,8 +348,7 @@

 # If the COPY is first, the VACUUM generates an error, so we vacuum first
 echo "vacuuming template1"
-echo "vacuum" | postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
-       grep -v "^DEBUG:"
+echo "vacuum" | postgres -o /dev/null -F -Q -D$PGDATA template1 > /dev/null

 echo "COPY pg_shadow TO '$PGDATA/pg_pwd' USING DELIMITERS '\\t'" |\
        postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
@@ -383,6 +382,6 @@
 echo "loading pg_description"
 echo "copy pg_description from '$TEMPLATE_DESCR'" | postgres -F -Q -D$PGDATA te
mplate1 > /dev/null
 echo "copy pg_description from '$GLOBAL_DESCR'" | postgres -F -Q -D$PGDATA temp
late1 > /dev/null
-echo "vacuum analyze" | postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
+echo "vacuum analyze" | postgres -o /dev/null -F -Q -D$PGDATA template1 > /dev/
null |\
        grep -v "^DEBUG:"

Any help appreciated !

Reply-To set, since I haven't subscribed to your list.

--
Andreas Klemm   http://www.FreeBSD.ORG/~andreas
powered by ,,symmetric multiprocessor FreeBSD''

On Sun, 22 Mar 1998, Andreas Klemm wrote:

> Hi !
>
> Please help, the FreeBSD port update to version 6.3 drives me to
> despair ;-)
>
> --with-tcl configure switch doesn't work to build libpgtcl
>
> I only get it to work if I set in the make environment
>
>     USE_TCL=true

This has been fixed in the upcoming patch...I'm going to be bundling up a
postgresql-v6.3.1.tar.gz and postgresql-v6.3-v6.3.1.gz file this
afternoon...

> Another problem: on FreeBSD-current (and I think this is true on
> FreeBSD-2.2.x as well)

    Odd, since my testing has always been done on FreeBSD-current, and
I've already done a couple of builds under FreeBSD 2.2-STABLE :(

    But, looking at your patches for initdb.sh, I've applied them, as
I think that it provides a non-shell dependent solution, which I think is
cleaner...

    I'm just about to do a build on my system, to make sure
everythingis okay before I bundle...but the new version should be up (and
announced) later this evening...

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org           secondary: scrappy@{freebsd|postgresql}.org

On Sun, 22 Mar 1998, Andreas Klemm wrote:

> Hi !
>
> Please help, the FreeBSD port update to version 6.3 drives me to
> despair ;-)
>
> --with-tcl configure switch doesn't work to build libpgtcl
>
> I only get it to work if I set in the make environment
>
>     USE_TCL=true

This has been fixed in the upcoming patch...I'm going to be bundling up a
postgresql-v6.3.1.tar.gz and postgresql-v6.3-v6.3.1.gz file this
afternoon...

> Another problem: on FreeBSD-current (and I think this is true on
> FreeBSD-2.2.x as well)

    Odd, since my testing has always been done on FreeBSD-current, and
I've already done a couple of builds under FreeBSD 2.2-STABLE :(

    But, looking at your patches for initdb.sh, I've applied them, as
I think that it provides a non-shell dependent solution, which I think is
cleaner...

    I'm just about to do a build on my system, to make sure
everythingis okay before I bundle...but the new version should be up (and
announced) later this evening...

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org           secondary: scrappy@{freebsd|postgresql}.org

I just updated the ftp site such that there is a postgresql-6.3.1.tar.gz
file, which corrects alot of post-release bugs that cropped up, including
those you listed below...

Let me know if this help...it hasn't been officially released yet, but,
unless anyone comes up with an critical problems with it, that tar file
will be what is announced tomorrow morning...



On Sun, 22 Mar 1998, Andreas Klemm wrote:

> Hi !
>
> Please help, the FreeBSD port update to version 6.3 drives me to
> despair ;-)
>
> --with-tcl configure switch doesn't work to build libpgtcl
>
> I only get it to work if I set in the make environment
>
>     USE_TCL=true
>
> configure --with-tcl doesn't seem to set USE_TCL=true in
> src/Makefile.global, so
>     src/interfaces/Makefile,
> which includes
>     src/Makefile.global
> fails, since USE_TCL isn't set to true. Workaround was to explicitely
> set USE_TCL=true when starting the make.
>
> Another problem: on FreeBSD-current (and I think this is true on
> FreeBSD-2.2.x as well)
>
> The make install fails when running the initdb script.
>
> Adding template1 database to pg_database...
> Running: postgres -boot -C -F -D/usr/local/pgsql/data -Q template1 < /tmp/create
> .13645
>
> vacuuming template1
> creating public pg_user view
> loading pg_description
> *** Error code 1
>
> The Postgresql 6.2 version needed a patch so that this works
> --- bin/initdb/initdb.sh.orig   Wed Apr  2 07:09:13 1997
> +++ bin/initdb/initdb.sh        Wed Apr  2 07:09:33 1997
> @@ -344,6 +344,5 @@
>
>  if [ $debug -eq 0 ]; then
>      echo "vacuuming template1"
> -    echo "vacuum" | postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
> -        grep -v "^DEBUG:"
> +    echo "vacuum" | postgres -o /dev/null -F -Q -D$PGDATA template1 > /dev/null
>
>  fi
>
> I enhanced this patch and now use
>     "-o /dev/null"
> and avoid
>     "2>&1"
>
> several lines below where similar commands can be found, but this
> doesn't solve the problem.
>
> --- bin/initdb/initdb.sh.orig   Sun Mar 22 12:47:51 1998
> +++ bin/initdb/initdb.sh        Sun Mar 22 13:06:36 1998
> @@ -348,8 +348,7 @@
>
>  # If the COPY is first, the VACUUM generates an error, so we vacuum first
>  echo "vacuuming template1"
> -echo "vacuum" | postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
> -       grep -v "^DEBUG:"
> +echo "vacuum" | postgres -o /dev/null -F -Q -D$PGDATA template1 > /dev/null
>
>  echo "COPY pg_shadow TO '$PGDATA/pg_pwd' USING DELIMITERS '\\t'" |\
>         postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
> @@ -383,6 +382,6 @@
>  echo "loading pg_description"
>  echo "copy pg_description from '$TEMPLATE_DESCR'" | postgres -F -Q -D$PGDATA te
> mplate1 > /dev/null
>  echo "copy pg_description from '$GLOBAL_DESCR'" | postgres -F -Q -D$PGDATA temp
> late1 > /dev/null
> -echo "vacuum analyze" | postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
> +echo "vacuum analyze" | postgres -o /dev/null -F -Q -D$PGDATA template1 > /dev/
> null |\
>         grep -v "^DEBUG:"
>
> Any help appreciated !
>
> Reply-To set, since I haven't subscribed to your list.
>
> --
> Andreas Klemm   http://www.FreeBSD.ORG/~andreas
> powered by ,,symmetric multiprocessor FreeBSD''
>

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org           secondary: scrappy@{freebsd|postgresql}.org

>
> Hi !
>
> Please help, the FreeBSD port update to version 6.3 drives me to
> despair ;-)
>
> --with-tcl configure switch doesn't work to build libpgtcl
>
> I only get it to work if I set in the make environment
>
>     USE_TCL=true
>
> configure --with-tcl doesn't seem to set USE_TCL=true in
> src/Makefile.global, so
>     src/interfaces/Makefile,
> which includes
>     src/Makefile.global
> fails, since USE_TCL isn't set to true. Workaround was to explicitely
> set USE_TCL=true when starting the make.

We are working on a patch for this.

>
> Another problem: on FreeBSD-current (and I think this is true on
> FreeBSD-2.2.x as well)
>
> The make install fails when running the initdb script.
>
> Adding template1 database to pg_database...
> Running: postgres -boot -C -F -D/usr/local/pgsql/data -Q template1 < /tmp/create
> .13645
>
> vacuuming template1
> creating public pg_user view
> loading pg_description
> *** Error code 1
>
> The Postgresql 6.2 version needed a patch so that this works
> --- bin/initdb/initdb.sh.orig   Wed Apr  2 07:09:13 1997
> +++ bin/initdb/initdb.sh        Wed Apr  2 07:09:33 1997
> @@ -344,6 +344,5 @@
>
>  if [ $debug -eq 0 ]; then
>      echo "vacuuming template1"
> -    echo "vacuum" | postgres -F -Q -D$PGDATA template1 2>&1 > /dev/null |\
> -        grep -v "^DEBUG:"
> +    echo "vacuum" | postgres -o /dev/null -F -Q -D$PGDATA template1 > /dev/null
>
>  fi
>

There must be something else different on your machine.  We all use it
without problem.  Are you running csh, or changed it somehow?

--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
  +  If your life is a hard drive,     |  (610) 353-9879(w)
  +  Christ can be your backup.        |  (610) 853-3000(h)

>     Odd, since my testing has always been done on FreeBSD-current, and
> I've already done a couple of builds under FreeBSD 2.2-STABLE :(
>
>     But, looking at your patches for initdb.sh, I've applied them, as
> I think that it provides a non-shell dependent solution, which I think is
> cleaner...
>
>     I'm just about to do a build on my system, to make sure
> everythingis okay before I bundle...but the new version should be up (and
> announced) later this evening...

I recommend against applying the patch.

First, it is very late in the patch game to be applying an initdb patch.
Second, the original code should work fine, and the top of the file
says #/bin/sh, so we don't need a shell-independent solution.  The rest
of the file uses /bin/sh commands so it will simply not work with
another shell.  I suggest the original user has some problem with their
stderr handling in their OS configuration.


--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
  +  If your life is a hard drive,     |  (610) 353-9879(w)
  +  Christ can be your backup.        |  (610) 853-3000(h)

>     Odd, since my testing has always been done on FreeBSD-current, and
> I've already done a couple of builds under FreeBSD 2.2-STABLE :(
>
>     But, looking at your patches for initdb.sh, I've applied them, as
> I think that it provides a non-shell dependent solution, which I think is
> cleaner...
>
>     I'm just about to do a build on my system, to make sure
> everythingis okay before I bundle...but the new version should be up (and
> announced) later this evening...

I recommend against applying the patch.

First, it is very late in the patch game to be applying an initdb patch.
Second, the original code should work fine, and the top of the file
says #/bin/sh, so we don't need a shell-independent solution.  The rest
of the file uses /bin/sh commands so it will simply not work with
another shell.  I suggest the original user has some problem with their
stderr handling in their OS configuration.


--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
  +  If your life is a hard drive,     |  (610) 353-9879(w)
  +  Christ can be your backup.        |  (610) 853-3000(h)

> I recommend against applying the patch.
>
> First, it is very late in the patch game to be applying an initdb patch.
> Second, the original code should work fine, and the top of the file
> says #/bin/sh, so we don't need a shell-independent solution.  The rest
> of the file uses /bin/sh commands so it will simply not work with
> another shell.  I suggest the original user has some problem with their
> stderr handling in their OS configuration.

Maybe -o is better.  Marc will decide.

--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
  +  If your life is a hard drive,     |  (610) 353-9879(w)
  +  Christ can be your backup.        |  (610) 853-3000(h)

BTW, what tcl tk version do you recommend ?
80 or 81 ?

--
Andreas Klemm   http://www.FreeBSD.ORG/~andreas
powered by ,,symmetric multiprocessor FreeBSD''

On Sun, Mar 22, 1998 at 04:02:52PM -0400, The Hermit Hacker wrote:
>
>
> I just updated the ftp site such that there is a postgresql-6.3.1.tar.gz
> file, which corrects alot of post-release bugs that cropped up, including
> those you listed below...
>
> Let me know if this help...it hasn't been officially released yet, but,
> unless anyone comes up with an critical problems with it, that tar file
> will be what is announced tomorrow morning...

Thanks, you were blindingly fast ;-)

Then maybe I will be commit ready if Satoshi agrees.

    Andreas ///

--
Andreas Klemm   http://www.FreeBSD.ORG/~andreas
powered by ,,symmetric multiprocessor FreeBSD''

On Sun, 22 Mar 1998, Andreas Klemm wrote:

> On Sun, Mar 22, 1998 at 04:02:52PM -0400, The Hermit Hacker wrote:
> >
> >
> > I just updated the ftp site such that there is a postgresql-6.3.1.tar.gz
> > file, which corrects alot of post-release bugs that cropped up, including
> > those you listed below...
> >
> > Let me know if this help...it hasn't been officially released yet, but,
> > unless anyone comes up with an critical problems with it, that tar file
> > will be what is announced tomorrow morning...
>
> Thanks, you were blindingly fast ;-)

    I wish I knew enough about building the ports (And had the time to
learn *sigh*)...then I could save you even more time :(

> > Thanks, you were blindingly fast ;-)
>
>     I wish I knew enough about building the ports (And had the time to
> learn *sigh*)...then I could save you even more time :(

Yeah, but don't desperate ;-)

--
Andreas Klemm   http://www.FreeBSD.ORG/~andreas
powered by ,,symmetric multiprocessor FreeBSD''