Hi,
Thanks. The build for 8.3RC1 worked and all regression tests were passed.
Regards
J6M
----- Original Message -----
From: "Tom Lane" <tgl@sss.pgh.pa.us>
To: "J6M" <j6m@adm.estp.fr>
Cc: <pgsql-ports@postgresql.org>
Sent: Tuesday, January 08, 2008 2:20 AM
Subject: Re: [PORTS] Building 8.3 beta 4 stops on (at least) 2 errors
> "J6M" <j6m@adm.estp.fr> writes:
>> xlc -O2 -qmaxmem=16384 -qsrcmsg -qlonglong -qnoansialias -I../../../../src/include
>> -c -o tsquery.o tsquery.c
>> 108 | PT_CLOSE = 5,
>> ....................a
>> a - 1506-275 (S) Unexpected text ',' encountered.
>
> Some compilers are picky about extra commas in enum lists, some aren't.
> I guess we have not currently got any in the buildfarm that are.
>
>> xlc -O2 -qmaxmem=16384 -qsrcmsg -qlonglong -qnoansialias -I../include
>> -I../../../../src/interfaces/ecpg/include
>> -I../../../../src/interfaces/libpq -I../../../../src/include/utils -I../../../../src/include
>> -c -o informix.o informix.c
>> "/usr/include/stdio.h", line 434.12: 1506-343 (S) Redeclaration of
>> fgetpos64
>> differs from previous declaration on line 274 of "/usr/include/stdio.h".
>> "/usr/include/stdio.h", line 434.12: 1506-377 (I) The type "long long*"
>> of
>> parameter 2 differs from the previous type "long*".
>
> The problem here seems to be that informix.c was violating our coding
> rule that no system headers may be included before c.h (or in this case,
> postgres_fe.h). Please change the first few lines of the file to
> this ordering:
>
> #define POSTGRES_ECPG_INTERNAL
> #include "postgres_fe.h"
>
> #include <math.h>
> #include <ctype.h>
> #include <limits.h>
>
> #include <ecpgtype.h>
> ...
>
> and see if it doesn't work then.
>
> BTW, it seems we do not have any buildfarm members running xlc.
> If you can spare the cycles to build PG once a day or so, please
> consider joining the buildfarm so these types of portability issues
> can be caught sooner.
> http://www.pgbuildfarm.org/index.html
>
> regards, tom lane
