Excerpts from Tom Lane's message of mar jul 05 10:47:03 -0400 2011:
> Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> > Move Trigger and TriggerDesc structs out of rel.h into a new reltrigger.h
> > This lets us stop including rel.h into execnodes.h, which is a widely
> > used header.
>
> I'm confused why this patch added pg_am.h to predtest.c?
Because otherwise it fails to compile with this error:
gcc -O2 -Wall -Wmissing-prototypes -Wpointer-arith -Wdeclaration-after-statement -Wendif-labels -Wformat-security
-fno-strict-aliasing-fwrapv -g -I../../../../src/include -I/pgsql/source/HEAD/src/include -D_GNU_SOURCE
-I/usr/include/libxml2 -c -o predtest.o /pgsql/source/HEAD/src/backend/optimizer/util/predtest.c -MMD -MP -MF
.deps/predtest.Po
/pgsql/source/HEAD/src/backend/optimizer/util/predtest.c: In function ‘get_btree_test_op’:
/pgsql/source/HEAD/src/backend/optimizer/util/predtest.c:1661:32: error: ‘BTREE_AM_OID’ undeclared (first use in this
function)
/pgsql/source/HEAD/src/backend/optimizer/util/predtest.c:1661:32: note: each undeclared identifier is reported only
oncefor each function it appears in
make: *** [predtest.o] Error 1
Since that symbol is defined in pg_am.h, I thought the most convenient
fix was to include just that file. I could, of course, have included
the whole of rel.h but that seemed a bit pointless. The relevant code
is actually dealing with opfamilies (pg_amop.h is already being
included):
/* Now search the opfamilies */for (i = 0; i < catlist->n_members; i++){ HeapTuple pred_tuple =
&catlist->members[i]->tuple; Form_pg_amop pred_form = (Form_pg_amop) GETSTRUCT(pred_tuple); HeapTuple
clause_tuple;
/* Must be btree */ if (pred_form->amopmethod != BTREE_AM_OID) continue;
(Of course, the reason this didn't fail previously is because rel.h
includes pg_am.h).
--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
