[Python-checkins] r46962 - in python/trunk: Lib/sqlite3/test/hooks.py Lib/sqlite3/test/userfunctions.py Modules/_sqlite/connection.c

[Neal Norwitz]

On 6/15/06, Gerhard Häring <gh at ghaering.de> wrote:
> Neal Norwitz wrote:
> > I forgot to mention that all the builbots (even one on 3.1.3 or is
> > that my laptop) pass with the change I checked in.  probably rev
> > 26964.
>
> Ok, but I believe they passed all tests before your checkin, too.

Right, the difference is that they passed before the checkin because
the code wasn't exercised if the version was too early.  They passed
after the checkin with the version checks removed, so the code was
executed.

> > On 6/14/06, Neal Norwitz <nnorwitz at gmail.com> wrote:
> >> Can you take a look at my checkin?  It only added 2 lines and the
> >> tests pass after removing the version checks.  I don't know if there's
> >> a good reason to *not* do the fix I did.
>
> Doing it this way is a possible approach that should work ok, and it's
> probably a good one, too. It is functionally different from the real
> thing (sqlite3_result_error) though, because the query in progress is
> not aborted.
>
> If you want to keep your checkin, please restore this change though:
>
> http://svn.python.org/view/python/trunk/Lib/sqlite3/test/hooks.py?rev=46964&r1=46962&r2=46964

Sure, I'll re-revert that change. (or will it be a re-re-revert?)

> it addresses a different problem that I haven't fully debugged yet. All
> I know is that without it, the test suite will crash for SQLite 3.0.8
> and not for 3.2.1 and later. It's most likely another SQLite bug of some
> sort.

It didn't crash for 3.1.3 on Mac OS X 10.3.  I didn't pay close
attention and verify the test was run though.  I couldn't find an
earlier version.

Please try to talk to Anthony so we can figure out how to proceed with
2.5 and if anything needs to be changed.  Probably the biggest concern
is that in order to support this properly, APIs will need to change.

n