]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/commit
Repair _execute_scalar for WITH_UNICODE mode
authorMike Bayer <mike_mp@zzzcomputing.com>
Mon, 13 Mar 2017 16:27:51 +0000 (12:27 -0400)
committerMike Bayer <mike_mp@zzzcomputing.com>
Mon, 13 Mar 2017 18:36:45 +0000 (14:36 -0400)
commitd2787feb228612d691c5992d61c2529408c36765
treeac0c8c3fa6179ae3598d885ccb8ebe0b9d2498f4
parent1a555a550869ca9e0602496a9f28fc7d2373dc3a
Repair _execute_scalar for WITH_UNICODE mode

cx_Oracle 5.3 seems to code this flag ON now, so
remove the warning and ensure WITH_UNICODE handling works.

Additionally, the XE setup on jenkins is having more
problems here, in particular low-connections mode is
causing cx_Oracle to fail more frequently now.  Turning
off low-connections fixes those but then we get the
TNS errors, so adding an emergency "retry" flag that
is not yet a feature available to users.  Real world
applications are not dropping/creating thousands of
tables the way our test suite is.

Change-Id: Ie95b0e697276c404d3264c2e624e870463d966d6
Fixes: #3937
doc/build/changelog/changelog_10.rst
lib/sqlalchemy/dialects/oracle/cx_oracle.py
lib/sqlalchemy/testing/provision.py
test/dialect/test_oracle.py
tox_1_1.ini