PostgreSQL 9.5 fails compiling against 2.2 in twkb

[robe]

Yikes — can't have this. Started happening with pull this weekend.

/lib/jenkins/workspace/pg/rel/pg9.5w64/include/postgresql/internal -D_GNU_SOURCE   -c -o lwgeom_inout.o lwgeom_inout.c
lwgeom_inout.c: In function ‘TWKBFromLWGEOMArray’:
lwgeom_inout.c:538: error: too few arguments to function ‘array_create_iterator’
lwgeom_inout.c:539: error: too few arguments to function ‘array_create_iterator’
make[1]: *** [lwgeom_inout.o] Error 1
make[1]: Leaving directory `/var/lib/jenkins/workspace/postgis/regress_pgdev/branches/2.2/postgis'

I suspect it was this commit of Tom's that broke our build:

​https://github.com/postgres/postgres/commit/1dc5ebc9077ab742079ce5dac9a6664248d42916

[robe]

Put in PostgreSQL bug ticket here:

​http://www.postgresql.org/message-id/20150518010719.1029.17180@wrigleys.postgresql.org

[robe]

Note from Peter Geoghegan for completeness - guess I got the commit wrong and should be: ​https://github.com/postgres/postgres/commit/13dbc7a8

though something doesn't seem right. That commit was done in March and I'm pretty sure debbie has been pulling the tar balls and rebuilding and we've even had regression issues since then. So why it's failing now is very odd.

---

On Sun, May 17, 2015 at 6:07 PM, wrote:

I suspect the problem commit is this: ​https://github.com/postgres/postgres/commit/1dc5ebc9077ab742079ce5dac9 a6664248d42916

Looks like commit 13dbc7a8 to me.

It is safe to just add a third argument to that function call, NULL. That's all that was done for the array_create_iterator() call in pl_exec.c by commit 13dbc7a8.

— Peter Geoghegan

---

[pramsey]

Fixed at r13524