The patches can be found here: https://anonscm.debian.org/cgit/pkg-mysql/mariadb-10.3.git/tree/debian/patches

• scripts_mysql_create_system_tables_no_test.patch
  Yes, we should do something like this along with https://github.com/MariaDB/server/pull/536
• scripts_mysqld_safe.sh_signals.patch
  Probably.
• scripts_mysql_install_db.sh_no_test.patch
  Yes, we should do something like this along with https://github.com/MariaDB/server/pull/536
• mysql-test__db_test.patch
  Yes, we should do something like this along with https://github.com/MariaDB/server/pull/536
• replace_dash_with_bash_mbug675185.patch
  Yes.
• extend_default_test_timeout_for_tokudb.patch
  Does it really take 9 hours to run the whole test suite?
• innodb_simulate_comp_test_speedup.patch
  Yes, looks nice.
• mysqld_multi_confd.patch
  Probably, we need to make sure it works well with RPM/tarball.
• mysqld_multi.server_lsb-header.patch
  This looks Debian specific. Do we still need this script?
• mips-groonga-atomic.patch
  This is most probably fixed.
• mips-connect-unaligned.patch
  The problem is valid, but we already have routines for loading/storing unalgined data.
• mips-machine.patch
  Don't we have this since quite a while already?
• hurd_socket.patch
  Yes.
• armhf_mroonga_storage_fail.patch
  Why?
• remove-systemd-obsolete-target.patch
  Yes, likely. IIRC there was PR fixing it.
• kFreeBSD-gettid.patch
  Yes, but the patch seem to be wrong: shouldn't it do "return thr_self()" instead of "thr_self(&lwpid)?"
• 0018-Fix-the-awk-path-on-Debian.patch
  Yes, but we just need to remove "#!/usr/bin/awk".
• mytop-merge_src:mytop_improvements.patch
  Probably.
• Add_default_ExecStartPre_to_mariadb@.service.patch
  Yes, it is likely a bug.
• 0021-Revert-to-using-system-pcre-library.patch
  Yes, but we need to check why did we bondle it in first place.
• 0025-Change-the-default-optimization-from-O3-to-O2-in-mys.patch
  Why?