Thread: Unable to build on Solaris 2.6
Hi,
I am unable to build PostgreSQL 7.1.2 on Solaris 2.6.
Below are some info as to versions of the tools used and error messages.
Can anybody make something out of this?
Thank you
Peter
MAKE VERSION:
GNU Make version 3.77, by Richard Stallman and Roland McGrath.
GCC VERSION:
gcc version egcs-2.91.66 19990314 (egcs-1.1.2 release)
JUST THE BEGINNING OF A LONG OUTPUT OF ERROR MESSAGES:
[ws6709] /tmp/postgresql-7.1.2$ gmake
gmake -C doc all
gmake[1]: Entering directory `/tmp/postgresql-7.1.2/doc'
gmake[1]: Nothing to be done for `all'.
gmake[1]: Leaving directory `/tmp/postgresql-7.1.2/doc'
gmake -C src all
gmake[1]: Entering directory `/tmp/postgresql-7.1.2/src'
gmake -C backend all
gmake[2]: Entering directory `/tmp/postgresql-7.1.2/src/backend'
gmake -C access all
gmake[3]: Entering directory `/tmp/postgresql-7.1.2/src/backend/access'
gmake -C common SUBSYS.o
gmake[4]: Entering directory
`/tmp/postgresql-7.1.2/src/backend/access/common'
gcc -Wall -Wmissing-prototypes -Wmissing-declarations
-I../../../../src/include -c heaptuple
.c -o heaptuple.o
In file included from ../../../../src/include/access/tupdesc.h:18,
from ../../../../src/include/utils/rel.h:18,
from ../../../../src/include/storage/bufmgr.h:21,
from ../../../../src/include/storage/bufpage.h:18,
from ../../../../src/include/access/htup.h:17,
from ../../../../src/include/access/heapam.h:18,
from heaptuple.c:23:
../../../../src/include/catalog/pg_attribute.h:43: syntax error before `{'
../../../../src/include/catalog/pg_attribute.h:156: parse error before `}'
../../../../src/include/catalog/pg_attribute.h:156: warning: type defaults
to `int' in declarat
ion of `FormData_pg_attribute'
../../../../src/include/catalog/pg_attribute.h:156: warning: data definition
has no type or sto
rage class
../../../../src/include/catalog/pg_attribute.h:170: parse error before `*'
../../../../src/include/catalog/pg_attribute.h:170: warning: type defaults
to `int' in declarat
ion of `Form_pg_attribute'
../../../../src/include/catalog/pg_attribute.h:170: warning: data definition
has no type or sto
rage class
In file included from ../../../../src/include/utils/rel.h:18,
Kovács Péter writes:
> gmake[3]: Entering directory `/tmp/postgresql-7.1.2/src/backend/access'
> gmake -C common SUBSYS.o
> gmake[4]: Entering directory
> `/tmp/postgresql-7.1.2/src/backend/access/common'
> gcc -Wall -Wmissing-prototypes -Wmissing-declarations
> -I../../../../src/include -c heaptuple
> .c -o heaptuple.o
> In file included from ../../../../src/include/access/tupdesc.h:18,
> from ../../../../src/include/utils/rel.h:18,
> from ../../../../src/include/storage/bufmgr.h:21,
> from ../../../../src/include/storage/bufpage.h:18,
> from ../../../../src/include/access/htup.h:17,
> from ../../../../src/include/access/heapam.h:18,
> from heaptuple.c:23:
> ../../../../src/include/catalog/pg_attribute.h:43: syntax error before `{'
> ../../../../src/include/catalog/pg_attribute.h:156: parse error before `}'
> ../../../../src/include/catalog/pg_attribute.h:156: warning: type defaults
> to `int' in declaration of `FormData_pg_attribute'
There's some macro trickery going on there, so it's hard to tell what it's
really seeing. Try this:
$ cd src/backend/access/common
$ make heaptuple.o CFLAGS=-save-temps
Then look at the resulting (I hope) heaptuple.i and look at what's
really happening at line 42 (search for pg_attribute should find it). It
should look similar to this:
# 42 "../../../../src/include/catalog/pg_attribute.h" 1
typedef struct FormData_pg_attribute
{
Oid attrelid;
NameData attname;
# 55 "../../../../src/include/catalog/pg_attribute.h"
Oid atttypid;
There's the "{" it's talking about.
--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter
Peter,
Thanks a lot.
The generated definition looked like:
typedef struct FormData_ pg_attribute
{
Oid attrelid;
NameData attname;
Oid atttypid;
Apparently, the configure script did not correctly recognize the stringizing
capibality of the preprocessor. (Should I report this as a bug in
autoconf???) Defining the HAVE_STRINGIZING macro in config.h by hand seemed
to solve the problem and the compilation completed without error (though
with lots of "<type>_{MIN | MAX} redefined" warnings).
Thank you again.
Peter
-----Original Message-----
From: Peter Eisentraut [mailto:peter_e@gmx.net]
Sent: Monday, July 16, 2001 11:56 PM
To: Kovács Péter
Cc: pgsql-ports@postgresql.org
Subject: Re: [PORTS] Unable to build on Solaris 2.6
Kovács Péter writes:
> gmake[3]: Entering directory `/tmp/postgresql-7.1.2/src/backend/access'
> gmake -C common SUBSYS.o
> gmake[4]: Entering directory
> `/tmp/postgresql-7.1.2/src/backend/access/common'
> gcc -Wall -Wmissing-prototypes -Wmissing-declarations
> -I../../../../src/include -c heaptuple
> .c -o heaptuple.o
> In file included from ../../../../src/include/access/tupdesc.h:18,
> from ../../../../src/include/utils/rel.h:18,
> from ../../../../src/include/storage/bufmgr.h:21,
> from ../../../../src/include/storage/bufpage.h:18,
> from ../../../../src/include/access/htup.h:17,
> from ../../../../src/include/access/heapam.h:18,
> from heaptuple.c:23:
> ../../../../src/include/catalog/pg_attribute.h:43: syntax error before `{'
> ../../../../src/include/catalog/pg_attribute.h:156: parse error before `}'
> ../../../../src/include/catalog/pg_attribute.h:156: warning: type defaults
> to `int' in declaration of `FormData_pg_attribute'
There's some macro trickery going on there, so it's hard to tell what it's
really seeing. Try this:
$ cd src/backend/access/common
$ make heaptuple.o CFLAGS=-save-temps
Then look at the resulting (I hope) heaptuple.i and look at what's
really happening at line 42 (search for pg_attribute should find it). It
should look similar to this:
# 42 "../../../../src/include/catalog/pg_attribute.h" 1
typedef struct FormData_pg_attribute
{
Oid attrelid;
NameData attname;
# 55 "../../../../src/include/catalog/pg_attribute.h"
Oid atttypid;
There's the "{" it's talking about.
--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter
Kovács Péter writes: > Apparently, the configure script did not correctly recognize the stringizing > capibality of the preprocessor. (Should I report this as a bug in > autoconf???) Hmm. I've never seen this test give a wrong result. Can you check *why* it does that? -- Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter