- 23 Apr, 2004 2 commits
-
-
Tim Peters authored
fiddly code trying to prevent OverflowError.
-
Tim Peters authored
conversion (this is an --all test, so the original committer probably didn't run it). Introduced a trivial new module (loglevels.py) to centralize creating the custom log levels ZODB/ZEO want to use.
-
- 22 Apr, 2004 4 commits
-
-
Gintautas Miliauskas authored
processing the messages because the format might have changed a little bit.
-
Fred Drake authored
-
Fred Drake authored
(the tests themselves still aren't being run; we need decent save/restore behavior from the logging package to do this properly)
-
Fred Drake authored
logging directly
-
- 20 Apr, 2004 1 commit
-
-
Fred Drake authored
later
-
- 19 Apr, 2004 5 commits
-
-
Tim Peters authored
didn't recognize its version number. As a result, 3.3a3 went out with ZODB/__init__.py claiming it was version 3.3a2. Repaired that, but by changing what release.py looks for.
-
Tim Peters authored
-
Tim Peters authored
-
Tim Peters authored
cause of the: TypeError: unpack() argument 2 must be string or read-only buffer, not None failure often seen in overnight testrunner reports now, coming from the checkConcurrentUpdatesInVersions test. If the hypothesis is correct, we'll see a failure of the new assert end is not None instead.
-
Barry Warsaw authored
returns a datetime.timedelta instead of a float. Added test case and documentation.
-
- 17 Apr, 2004 4 commits
-
-
Gintautas Miliauskas authored
-
Gintautas Miliauskas authored
-
Gintautas Miliauskas authored
-
Gintautas Miliauskas authored
-
- 16 Apr, 2004 24 commits
-
-
Jeremy Hylton authored
-
Tim Peters authored
-
Tim Peters authored
-
Jeremy Hylton authored
-
Tim Peters authored
-
Jeremy Hylton authored
-
Jeremy Hylton authored
-
Tim Peters authored
-
Tim Peters authored
thinks they live.
-
Jeremy Hylton authored
-
Jeremy Hylton authored
The old code used itertools.chain(), which didn't work as expected. If the first of the two iterator grew after the chain had started consuming the second iterator, the new element(s) in the first iterator would never be consumed. Fix this bug by eliminating the need for chaining. Process all the _added_during_commit objects at the end, each with its own ObjectWriter instance.
-
Jeremy Hylton authored
-
Jeremy Hylton authored
Add explanation for __getattr__ and friends.
-
Jeremy Hylton authored
-
Tim Peters authored
raised when an operation on a Connection can't be performed because the Connection is in "a wrong state". Connection.close() now raises this exception if the connection is currently joined to a transaction (most obviously, if the transaction the connection is joined to has modified objects waiting for commit or abort). Also changed other appropriate instances of RuntimeError to raise ConnectionStateError instead (e.g., trying to load an object from a closed connection).
-
Jeremy Hylton authored
-
Tim Peters authored
-
Jeremy Hylton authored
-
Jeremy Hylton authored
-
Jeremy Hylton authored
-
Jeremy Hylton authored
Replace use of type('') with basestring. The code didn't call setLocalTransaction(), so getTransaction() always did the same thing as transaction.get().
-
Jeremy Hylton authored
-
Jeremy Hylton authored
-
Jeremy Hylton authored
-