summaryrefslogtreecommitdiff
path: root/alembic/ddl/mysql.py
diff options
context:
space:
mode:
authorMike Bayer <mike_mp@zzzcomputing.com>2014-11-30 18:02:57 -0500
committerMike Bayer <mike_mp@zzzcomputing.com>2014-11-30 18:02:57 -0500
commit96bbcf2542959b5b0dc99bbfb0db9a1db7072bc5 (patch)
tree1bef6c1a3ec78296b0d7b49491de0fdc3008a9e4 /alembic/ddl/mysql.py
parentc4409d67985fca04efdad02da8d3a7b8bc7cdc6f (diff)
downloadalembic-96bbcf2542959b5b0dc99bbfb0db9a1db7072bc5.tar.gz
- refactor the FK merge a bit
- getting at attributes of FKs varies a bit on SQLA versions, so implement an _fk_spec() called for all FK inspection - to enable include_object() filters and allow the FK constraint code to flow like that of indexes/uniques, change the approach so that we deal with an _fk_constraint_sig() object again which contains the real ForeignKeyConstraint() within; we need this anyway for include_object, but also allows us to use the standard "drop_constraint" call for rendering. - enhance tests in test_autogen_fks to support real FK databases like Postgresql, MySQL, add in InnoDB flags and ensure that FKs refer to real primary key constraints for PG support - implement and test include_object() support for FKs - inspectors all have get_foreign_keys(), no need to check - repair the drop_constraint call to quote the "type" and table name correctly, run all constraint drops through drop_constraint() for rendering - fix up schema identifiers for foreign key autogens
Diffstat (limited to 'alembic/ddl/mysql.py')
-rw-r--r--alembic/ddl/mysql.py1
1 files changed, 0 insertions, 1 deletions
diff --git a/alembic/ddl/mysql.py b/alembic/ddl/mysql.py
index aac8184..b93d29f 100644
--- a/alembic/ddl/mysql.py
+++ b/alembic/ddl/mysql.py
@@ -95,7 +95,6 @@ class MySQLImpl(DefaultImpl):
# TODO: if SQLA 1.0, make use of "duplicates_index"
# metadata
-
removed = set()
for idx in list(conn_indexes):
# MySQL puts implicit indexes on FK columns, even if