postgres83 fails to build +universal

[jeremyhu (Jeremy Huddleston Sequoia)]

/usr/bin/gcc-4.2 -no-cpp-precomp -O0 -ggdb3 -arch x86_64 -arch i386 -Wall -Wmissing-prototypes -Wpointer-arith -Winline -Wdeclaration-after-statement -Wendif-labels -fno-strict-aliasing -fwrapv -I../../../../src/include -I/opt/local/include -I/opt/local/include/ossp -I/opt/local/include/libxml2  -I/opt/local/include  -c -o heaptuple.o heaptuple.c
heaptuple.c: In function 'heap_fill_tuple':
heaptuple.c:250: error: duplicate case value
heaptuple.c:250: error: previously used here
heaptuple.c: In function 'DataFill':
heaptuple.c:383: error: duplicate case value
heaptuple.c:383: error: previously used here
lipo: can't figure out the architecture type of: /var/tmp//cceMu6DA.out
gnumake[4]: *** [heaptuple.o] Error 1
gnumake[3]: *** [common-recursive] Error 2
gnumake[2]: *** [access-recursive] Error 2
gnumake[1]: *** [all] Error 2
gnumake: *** [all] Error 2
shell command " cd "/opt/local/var/macports/build/_Users_jeremy_src_macports-trunk_dports_databases_postgresql83/work/postgresql-8.3.10" && /usr/bin/gnumake " returned error 2
Error: Target org.macports.build returned: shell command failed
Warning: the following items did not execute (for postgresql83): org.macports.destroot org.macports.build
Log for postgresql83 is at: /opt/local/var/macports/logs/_Users_jeremy_src_macports-trunk_dports_databases_postgresql83/main.log
Error: Unable to upgrade port: 1
To report a bug, see <http://guide.macports.org/#project.tickets>

Looks like the problem is in the macro for store_att_byval as that is what is expanded at the erroring lines. This macro is defined in src/include/access/tupmacs.h:

/*
 * store_att_byval is a partial inverse of fetch_att: store a given Datum
 * value into a tuple data area at the specified address.  However, it only
 * handles the byval case, because in typical usage the caller needs to
 * distinguish by-val and by-ref cases anyway, and so a do-it-all macro
 * wouldn't be convenient.
 */
#if SIZEOF_DATUM == 8

#define store_att_byval(T,newdatum,attlen) \
	do { \
		switch (attlen) \
		{ \
			case sizeof(char): \
				*(char *) (T) = DatumGetChar(newdatum); \
				break; \
			case sizeof(int16): \
				*(int16 *) (T) = DatumGetInt16(newdatum); \
				break; \
			case sizeof(int32): \
				*(int32 *) (T) = DatumGetInt32(newdatum); \
				break; \
			case sizeof(Datum): \
				*(Datum *) (T) = (newdatum); \
				break; \
			default: \
				elog(ERROR, "unsupported byval length: %d", \
					 (int) (attlen)); \
				break; \
		} \
	} while (0)
#else							/* SIZEOF_DATUM != 8 */

#define store_att_byval(T,newdatum,attlen) \
	do { \
		switch (attlen) \
		{ \
			case sizeof(char): \
				*(char *) (T) = DatumGetChar(newdatum); \
				break; \
			case sizeof(int16): \
				*(int16 *) (T) = DatumGetInt16(newdatum); \
				break; \
			case sizeof(int32): \
				*(int32 *) (T) = DatumGetInt32(newdatum); \
				break; \
			default: \
				elog(ERROR, "unsupported byval length: %d", \
					 (int) (attlen)); \
				break; \
		} \
	} while (0)
#endif   /* SIZEOF_DATUM == 8 */

It looks like SIZEOF_DATUM is probably getting set during configure based on just one of the architectures

[jmroot (Joshua Root)]

Should be essentially the same fix as for 8.4 in r63149.

[jmroot (Joshua Root)]

#14619