Don't change asyncpg's "char" codec

This codec was used to ensure the "pg_attribute.generated"
column comes back as a string and not bytes, matching how
other PG drivers treat this datatype.   However, this breaks
asyncpg's internal implementation of set_type_codec going forward
and the "char" datatype is actually a bytes in any case.
So at the moment it appears psycopg2/pg8000 are broken for mis-treatment
of the datatype and asyncpg is broken in that it was allowing
us to change a codec that it appears to rely upon internally.

Fixes: #5586
Change-Id: I937eba315904721aa4e2726b95432910a8affe5f

1 files changed, 8 insertions, 4 deletions

diff --git a/lib/sqlalchemy/dialects/postgresql/base.py b/lib/sqlalchemy/dialects/postgresql/base.py
index 5e81586b4..878693866 100644
--- a/lib/sqlalchemy/dialects/postgresql/base.py
+++ b/lib/sqlalchemy/dialects/postgresql/base.py
@@ -3423,10 +3423,14 @@ class PGDialect(default.DefaultDialect):
             )
             coltype = sqltypes.NULLTYPE
 
-        # If a zero byte (''), then not a generated column.
-        # Otherwise, s = stored. (Other values might be added in the future.)
-        if generated:
-            computed = dict(sqltext=default, persisted=generated == "s")
+        # If a zero byte or blank string depending on driver (is also absent
+        # for older PG versions), then not a generated column. Otherwise, s =
+        # stored. (Other values might be added in the future.)
+        #
+        if generated not in (None, "", b"\x00"):
+            computed = dict(
+                sqltext=default, persisted=generated in ("s", b"s")
+            )
             default = None
         else:
             computed = None