Обсуждение: Postgresql 7.4 fails to compile on Redhat Fedora Core 1 Linux

Hi all,

I just tried to compile Postgresql 7.4 on RHFC1 linux without success. I used
./configure --with-python --with-perl , then make.

I get an error while bootstrap.c is being compiled (see messages below).

gcc --version reports itself as
gcc (GCC) 3.3.2 20031022 (Red Hat Linux 3.3.2-1)

Maybe this is a gcc error? Any help would much be appreciated.

Regards
Marcel Gsteiger
Milprog AG
Switzerland

btw: There is an error on the 'subscribe to the list' link on http://archives.postgresql.org/pgsql-ports/ . One gets
theerror 

Majordomo Error

The domain "postgresql.o" is not supported!

..obviously the domain name should be postgresql.org instead.


error during compile:

gcc -O2 -fno-strict-aliasing -Wall -Wmissing-prototypes -Wmissing-declarations -I. -I../../../src/include -D_GNU_SOURCE
 -c -o bootstrap.o bootstrap.c 
bootstrap.c:118: error: `F_REGCLASSIN' undeclared here (not in a function)
bootstrap.c:118: error: initializer element is not constant
bootstrap.c:118: error: (near initialization for `Procid[8].inproc')
bootstrap.c:118: error: `F_REGCLASSOUT' undeclared here (not in a function)
bootstrap.c:118: error: initializer element is not constant
bootstrap.c:118: error: (near initialization for `Procid[8].outproc')
bootstrap.c:118: error: initializer element is not constant
bootstrap.c:118: error: (near initialization for `Procid[8]')
bootstrap.c:119: error: `F_REGTYPEIN' undeclared here (not in a function)
bootstrap.c:119: error: initializer element is not constant
bootstrap.c:119: error: (near initialization for `Procid[9].inproc')
bootstrap.c:119: error: `F_REGTYPEOUT' undeclared here (not in a function)
bootstrap.c:119: error: initializer element is not constant
bootstrap.c:119: error: (near initialization for `Procid[9].outproc')
bootstrap.c:119: error: initializer element is not constant
bootstrap.c:119: error: (near initialization for `Procid[9]')
bootstrap.c:120: error: initializer element is not constant
bootstrap.c:120: error: (near initialization for `Procid[10]')
bootstrap.c:121: error: initializer element is not constant
bootstrap.c:121: error: (near initialization for `Procid[11]')
bootstrap.c:122: error: initializer element is not constant
bootstrap.c:122: error: (near initialization for `Procid[12]')
bootstrap.c:123: error: initializer element is not constant
bootstrap.c:123: error: (near initialization for `Procid[13]')
bootstrap.c:124: error: initializer element is not constant
bootstrap.c:124: error: (near initialization for `Procid[14]')
bootstrap.c:125: error: initializer element is not constant
bootstrap.c:125: error: (near initialization for `Procid[15]')
bootstrap.c:126: error: initializer element is not constant
bootstrap.c:126: error: (near initialization for `Procid[16]')
bootstrap.c:127: error: initializer element is not constant
bootstrap.c:127: error: (near initialization for `Procid[17]')
bootstrap.c:128: error: initializer element is not constant
bootstrap.c:128: error: (near initialization for `Procid[18]')
make[3]: *** [bootstrap.o] Fehler 1
make[3]: Leaving directory `/usr/local/src/postgresql-7.4/src/backend/bootstrap'
make[2]: *** [bootstrap-recursive] Fehler 2
make[2]: Leaving directory `/usr/local/src/postgresql-7.4/src/backend'
make[1]: *** [all] Fehler 2
make[1]: Leaving directory `/usr/local/src/postgresql-7.4/src'
make: *** [all] Fehler 2

Marcel Gsteiger wrote:
> I just tried to compile Postgresql 7.4 on RHFC1 linux without
> success. I used ./configure --with-python --with-perl , then make.
>
> I get an error while bootstrap.c is being compiled (see messages
> below).

Can you send me the files config.log and src/backend/utils/fmgroids.h
(off list), please?  Something weird is going on here.  You're not the
first to report this.

Marcel Gsteiger wrote:
> Hi all,
>
> I just tried to compile Postgresql 7.4 on RHFC1 linux without
> success. I used ./configure --with-python --with-perl , then make.
>
> I get an error while bootstrap.c is being compiled (see messages
> below).

I'm lost on this.  A colleague just tried it on his RHFC1 machine, and I
went without problems.  I can't find any system in the errors in the
generated files.  It seems you will need to debug the problem yourself.
Try to take apart the script Gen_fmgrtab.sh in src/backend/utils.

Peter Eisentraut <peter_e@gmx.net> writes:
> I'm lost on this.  A colleague just tried it on his RHFC1 machine, and I
> went without problems.  I can't find any system in the errors in the
> generated files.

Did the errors match what the other person with this problem reported?
I couldn't see any pattern to those either; but it'd be helpful to
know if the failures are repeatable ...

            regards, tom lane

On Thu, 11 Dec 2003, Tom Lane wrote:

> Did the errors match what the other person with this problem reported?
> I couldn't see any pattern to those either; but it'd be helpful to
> know if the failures are repeatable ...

If you're referring to my message no.  I've been pulled off to another
project and will be for a few more days so haven't been able to do any
tracking.  Sorry.


Rod
--
    "Open Source Software - You usually get more than you pay for..."
     "Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL"

Tom Lane wrote:
> Peter Eisentraut <peter_e@gmx.net> writes:
> > I'm lost on this.  A colleague just tried it on his RHFC1 machine,
> > and I went without problems.  I can't find any system in the errors
> > in the generated files.
>
> Did the errors match what the other person with this problem
> reported?

There were no errors.  He said he's seldom seen a cleaner compile. :)

Peter Eisentraut <peter_e@gmx.net> writes:
> Tom Lane wrote:
>> Did the errors match what the other person with this problem
>> reported?

> There were no errors.  He said he's seldom seen a cleaner compile. :)

I'm sorry, I took you to mean that Marcel had sent you the bad
fmgroids.h and fmgrtab.c files off-list (as I recall you requested)
and you couldn't see any pattern in the omissions in them.  Which is
what I had to conclude after looking at the bad files sent me by the
previous complainant.  I still have those, if there's anything we can
compare them to.

There is obviously *something* weird going on on RHFC1, but it seems
like the first problem is to figure out how to reproduce it ...

            regards, tom lane

Marcel Gsteiger wrote:
> I could reproduce the error again after "make distclean".
> However, I could work around the problem as follows:
>  I tried "make distclean", " ./configure"  (without options) and
> "make". Surprisingly this worked without errors. After this, I could
> even do a "./configure --with-python" and "make" without any further
> errors.

I tried that and various combinations of switches and environments but
could not reproduce this.  The only difference between your system and
mine that I could tell from config.log is that you're using the smp
kernel.

Here are my versions of various involved software packages:

gcc-3.3.2-1
libgcc-3.3.2-1
bash-2.05b-31
make-3.79.1-18
gawk-3.1.3-3
sed-4.0.8-1

Also, try to see if you have a broken locale setting or something weird
in the environment variables.

Peter Eisentraut <peter_e@gmx.net> writes:
> I tried that and various combinations of switches and environments but
> could not reproduce this.

I find it striking that both complainants reported that the issue went
away on a fresh build attempt.  This seems to suggest that even on
configurations that can produce the problem, it only happens one time
in N, where we have little idea how large N is.

Marcel, could you try just doing "make distclean, configure, make"
ten or a hundred times overnight, and see if you get more failures?

            regards, tom lane