> >> "Harvey, Allan AC" <HarveyA@OneSteel.com> writes:
> >>> creating template1 database in
> /upgrades/postgres/postgresql-8.3.3/src/test/regress/./tmp_che
> ck/data/base/1 ... =: is not an identifier
>
> > A diff of postgres.bki on the SCO system to another that
> completed the regression
> > tests OK produced no differences.
>
> Huh ... so much for the easy theory. What you need to do
> next is figure
> out exactly where the failure is coming from. Try running initdb with
> --debug option (this will produce LOTS of stuff on stderr),
> and post the
> last hundred or so lines of debug output.
>
> regards, tom lane
>
This is mainly for the archives should someone else like to beat SCO into partial submission....
Regression testing does not work for reasons I don't know when attempting them from the makefiles.
ie make check.
BUT screen scrapping the executed commands and/or suggestions does work.
Here is what worked ( scrapped from bash history, hope I go it right ). Substitute you own installation path.
cd src/test/regress
rm -r /upgrades/postgres/postgresql-8.3.3/src/test/regress/./tmp_check/data
/upgrades/postgres/postgresql-8.3.3/src/test/regress/./tmp_check/install//usr/local/pgsql/bin/initdb -D /upgrades/postgres/postgresql-8.3.3/src/test/regress/./tmp_check/data -L /upgrades/postgres/postgresql-8.3.3/src/test/regress/./tmp_check/install/usr/local/pgsql/share --noclean --no-locale
/upgrades/postgres/postgresql-8.3.3/src/test/regress/./tmp_check/install//usr/local/pgsql/bin/postgres -D /upgrades/postgres/postgresql-8.3.3/src/test/regress/./tmp_check/data
In another shell...
cd src/test/regress
./pg_regress --psqldir=/upgrades/postgres/postgresql-8.3.3/src/test/regress/./tmp_check/install//usr/local/pgsql/bin --schedule=./serial_schedule --srcdir=/upgrades/postgres/postgresql-8.3.3/src/test/regress --multibyte=SQL_ASCII --load-language=plpgsql --host=localhost
All tests passed, save for those that use "Infinity" and "NaN" and variants on that theme. I can post regression.diffs
should anyone want them.
Regression testing is however the end of the story, getting it compiled was the main effort.
Summarised here:
Get GNU gawk and tar and substitute the SCO distributed versions.
Do not use the SCO Open server development environment ( export CC=cc; ./configure; etc etc ) as 64 bit is not supported an even more regression tests fail.
The Unixware and OpenServer Development Kit compiles nicely, but promptly core dumps with unsupported system calls in the compatibility module or some such.
Attempting to use the Skunkware gcc (2.95.3) fails in ./configure ( export CC=gcc ).
With the help of http://archives.postgresql.org/pgsql-bugs/2001-06/msg00141.php
and http://archives.postgresql.org/pgsql-ports/2008-06/msg00001.php
The correct options can be given to gcc.
src/makefiles/Makefile.sco is the wrong place, see link above,
/src/template/sco is the right place and I made it
#CC="$CC -b elf"
CC="$CC -melf"
The make worked which lead to the regression testing, see above.
Note UNIX sockets seem broken for get/setsockopt() so always connect over TCP/IP. ie --host=localhost seems to work around it.
Hope this helps someone else.
Now to the actuall upgrade.....
Allan
The material contained in this email may be confidential, privileged or copyrighted. If you are not the intended recipient, use, disclosure or copying of this information is prohibited. If you have received this document in error, please advise the sender and delete the document. Neither OneSteel nor the sender accept responsibility for any viruses contained in this email or any attachments.
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
No comments:
Post a Comment