This commit ports multimaster to Postgres 14:- removed tablespace-setup from mmts/Makefile  (according to commit 6c788d9f6aadb41d76a72d56149268371a7895ee);- add readOnlyTree flag to MtmProcessUtility()  (according to commit 7c337b6b527b7052e6a751f966d5734c56f668b5);- explicitly specify the hash function HASH_STRING  (according to commit b3817f5f774663d55931dd4fab9c5a94a15ae7ab);- refactor the use of EState  (according to commit 1375422c7826a2bf387be29895e961614f69de4b);- add 'include_out_arguments' to the function FuncnameGetCandidates()  (caccording to commit e56bce5d43789cce95d099554ae9593ada92b3b7);- use MyProc->xmin instead of MyPgXact->xmin in dmq.c  (according to commit 73487a60fc1063ba4b5178b69aee4ee210c182c4);- remove EE pooler support;- use SearchNamedReplicationSlot() instead of ReplicationSlotAcquire()

1) Remove the target installcheck, because it is used during make   installcheck-world.2) Remove placeholder for the target check, because it looks like everything is   already working..

Due to changing PROC->xmin type to pg_atomic_uint64.

…notests

For a build without assertions we get the following warning:src/resolver.c: In function ‘handle_response’:src/resolver.c:616:37: error: ‘gid’ may be used uninitialized in this function[-Werror=maybe-uninitialized]