Oracle Backend Updated

[Josh Juneau]

I've checked in all of the required code to make the Oracle backend function as expected.  I have also modified that code which was previously checked-in as to tune it a bit and make it easier to maintain/move forward.  I've been testing the Oracle backend for several weeks now and all of the required functionality appears to be in place.  I have not yet made any adjustments to enable Django 1.1 compatibility, but this code should be sufficient to release a 1.0 for use with the 1.0.X branch of Django.

Please let me know if you have any questions.

Thanks

Josh Juneau
june...@gmail.com
http://jj-blogger.blogspot.com
Twitter ID:  javajuneau

[Leo Soto M.]

On Thu, Oct 22, 2009 at 9:39 AM, Josh Juneau <june...@gmail.com> wrote:
> I've checked in all of the required code to make the Oracle backend function
> as expected.  I have also modified that code which was previously checked-in
> as to tune it a bit and make it easier to maintain/move forward.  I've been
> testing the Oracle backend for several weeks now and all of the required
> functionality appears to be in place.  I have not yet made any adjustments
> to enable Django 1.1 compatibility, but this code should be sufficient to
> release a 1.0 for use with the 1.0.X branch of Django.

Very cool!

I'm heading to a local conference now, but probably will have time at
Sunday to roll the 1.0 release.

Regards,
--
Leo Soto M.
http://blog.leosoto.com

[Jacob Fenwick]

I'll give a quick update as well.

I've checked out the mysql branch and started working on it again.

I've run tests for the mysql backend and the postgres backend against django 1.0.3 on Ubuntu. I've run these tests for django on jython 0.9 (the version it was one when the mysql branch was created), as well as for in the 1.0b version.

These are my results:

mysql_errors_jython_0.9.log
FAILED: (failures=40, errors=7)

mysql_errors_jython_1.0b.log
FAILED: (failures=40, errors=7)

postgres_errors_jython_0.9.log
FAILED: (failures=41, errors=19)

postgres_errors_jython_1.0b.log
FAILED: (failures=41, errors=19)

Just curious, do the number of failures and errors match your number of errors and failures for the mysql and postgres backends?

I've also tried to run the tests against Windows XP SP 3 but the tests never finish, even when left running for several days.

I started looking at the failures and errors that resulted from the test using the mysql backend and most of them are the same as the postgres failures and errors. Does this mean these are problems with Jython or problems with the backends?

I'm not sure how to proceed from this point forward and I'm looking for some advice.

Jacob

[Leo Soto M.]

On Fri, Oct 23, 2009 at 1:02 AM, Jacob Fenwick <jacob....@gmail.com> wrote:
> I'll give a quick update as well.
>
> I've checked out the mysql branch and started working on it again.
>
> I've run tests for the mysql backend and the postgres backend against django
> 1.0.3 on Ubuntu. I've run these tests for django on jython 0.9 (the version
> it was one when the mysql branch was created), as well as for in the 1.0b
> version.
>
> These are my results:
>
> mysql_errors_jython_0.9.log
> FAILED: (failures=40, errors=7)
>
> mysql_errors_jython_1.0b.log
> FAILED: (failures=40, errors=7)
>
> postgres_errors_jython_0.9.log
> FAILED: (failures=41, errors=19)
>
> postgres_errors_jython_1.0b.log
> FAILED: (failures=41, errors=19)

Can you also post the number of passing tests?

> Just curious, do the number of failures and errors match your number of
> errors and failures for the mysql and postgres backends?

I'll verify that on Sunday.

> I've also tried to run the tests against Windows XP SP 3 but the tests never
> finish, even when left running for several days.

Can you remind me what were the test that hanged the process?

> I started looking at the failures and errors that resulted from the test
> using the mysql backend and most of them are the same as the postgres
> failures and errors. Does this mean these are problems with Jython or
> problems with the backends?

Most of them are false positives (i.e, tests that are written in a way
that don't work with Jython even if the underlying feature does indeed
work in practice. A few Others are bugs in Django (there are a couple
of ticket open in their tracker), and others probably indicate Jython
bugs that we haven't addressed. I'll take a closer look if time
suffices.

> I'm not sure how to proceed from this point forward and I'm looking for some
> advice.

Just make sure the code is in good shape (you know, avoid duplicated
stuff, keep it readable, etc) and I'll merge it to trunk for 1.0.

[Jacob Fenwick]

On Fri, Oct 23, 2009 at 8:05 AM, Leo Soto M. <leo....@gmail.com> wrote:

On Fri, Oct 23, 2009 at 1:02 AM, Jacob Fenwick <jacob....@gmail.com> wrote:
> I'll give a quick update as well.
>
> I've checked out the mysql branch and started working on it again.
>
> I've run tests for the mysql backend and the postgres backend against django
> 1.0.3 on Ubuntu. I've run these tests for django on jython 0.9 (the version
> it was one when the mysql branch was created), as well as for in the 1.0b
> version.
>
> These are my results:
>
> mysql_errors_jython_0.9.log
> FAILED: (failures=40, errors=7)
>
> mysql_errors_jython_1.0b.log
> FAILED: (failures=40, errors=7)
>
> postgres_errors_jython_0.9.log
> FAILED: (failures=41, errors=19)
>
> postgres_errors_jython_1.0b.log
> FAILED: (failures=41, errors=19)

Can you also post the number of passing tests?

All 4 tests run 669 tests. I assume the tests that don't fail and don't produce an error pass.
 

> Just curious, do the number of failures and errors match your number of
> errors and failures for the mysql and postgres backends?

I'll verify that on Sunday.

> I've also tried to run the tests against Windows XP SP 3 but the tests never
> finish, even when left running for several days.

Can you remind me what were the test that hanged the process?

The test were it hangs on windows is:
--help is handled as a special case ...
 

> I started looking at the failures and errors that resulted from the test
> using the mysql backend and most of them are the same as the postgres
> failures and errors. Does this mean these are problems with Jython or
> problems with the backends?

Most of them are false positives (i.e, tests that are written in a way
that don't work with Jython even if the underlying feature does indeed
work in practice. A few Others are bugs in Django (there are a couple
of ticket open in their tracker), and others probably indicate Jython
bugs that we haven't addressed. I'll take a closer look if time
suffices.

If there's a list somewhere of all these bugs I could diff it with my bugs and try to attack any that are actual bugs in the mysql backend.
 

> I'm not sure how to proceed from this point forward and I'm looking for some
> advice.

Just make sure the code is in good shape (you know, avoid duplicated
stuff, keep it readable, etc) and I'll merge it to trunk for 1.0.

I will do a code review soon and do some manual testing to make sure it works.

Jacob

[Leo Soto M.]

I don't see any recent commit on the svn repository from you. Is it
working as is or you have some uncommited code on your side?

[Jacob Fenwick]

It's working as is.

Sorry it's taken me so long to get back to this. It really should have been committed a long time ago!

Jacob

[Leo Soto M.]

On Sun, Oct 25, 2009 at 8:40 PM, Jacob Fenwick <jacob....@gmail.com> wrote:
> It's working as is.
>
> Sorry it's taken me so long to get back to this. It really should have been
> committed a long time ago!

Thanks for your work -- It's already merged in on trunk!

I'm assuming that you will continue maintaining it in case we get bug
reports or feature requests :)

[Actually, I'm not a friend of "code ownership" as in "nobody can
touch the MySQL backend because it's Jacob's territory". But since you
are the most familiar with the code you will be able to improve/fix it
more quicky than the rest of us]

Regards, and congrats again for getting the backend to such usable state!