[ANN] PySQLite 1.0

[Gerhard Haering]

=============================
Announcement for PySQLite 1.0
=============================

I'm pleased to announce that PySQLite has reached version 1.0 now.

What the heck is it?
SQLite is a powerful embedded relational database engine. PySQLite makes it
available to Python developers via the Database API 2.0. If you want a
relational database, but not worry about installing a MySQL server, or
depending on a Microsoft JET engine, let alone their licensing and/or
cross-platform issues, (Py)SQLite is the solution for you.

This release features the following minor changes:

- "Fixed" bug #1039380. Now you can use subclasses of Unicode in your .execute()
calls. It is consistent with the current API. The new API in PySQLite 2.0 will
lose all this politically correct isinstance stuff, though.

- Added patch #989224: Build support for Darwin.

- Applied patch #992094: Don't start transaction on VACUUM or DETACH.

- Added note to docs about -DTHREADSAFE=1 being a good idea when compiling SQLite.


Future development
==================

PySQLite was built on SQLite 2.x, which is in maintenance-only mode now. The
same will go for PySQLite 1.0 from now on.

My development effort will concentrate on bringing SQLite3 to Python
developers, and creating a leightweight, DB-API conformant wrapper around it
that will have a better designed API than the old PySQLite. This API will be
more leightweight, strive for best performance, and be completely incompatible
with the old PySQLite.

At the same time, I will try to incrementally change the existing codebase, so
that it will compile against SQLite 3.x, and maybe try to convert it to the
vm-based API of SQLite instead of the callback-based one. This is for those
that have applications built around the old API.

Hamburg, 2004-10-05 Gerhard Häring

 

[Irmen de Jong]

Announcement for PySQLite 1.0
Yay!

- Added note to docs about -DTHREADSAFE=1 being a good idea when compiling SQLite.

Indeed, once I sorted out the multithreading issues that I had
in my server app, it has now been running reliably for 3+ weeks non-stop.
Who said that online web applications need a 'true' database backend
(if anyone is interested, the site is www.promozilla.nl)

Good luck with new Pysqlite for sqlite 3. It's interesting stuff.

--Irmen de Jong

 

[Alex Martelli]

=============================
Announcement for PySQLite 1.0
=============================

I'm pleased to announce that PySQLite has reached version 1.0 now.

Cheers and celebrations!!! It's always important to hear when a
developer deems his code mature and solid enough to risk the "1.0" label
-- and this goes double for the _excellent_ PySQLite, which was a pretty
solid and mature and _incredibly useful_ piece of code for quite a while
already. Bravo Gerhard!

I hope the DBAPI SIG's efforts to make a better DBAPI 3.0 aren't quite
stalled...? I think it's high time to standardize down to ONE format
for argument substitution and support _iteration_ on fetches rather than
having to either fetch one, or a list... that's a minimum set on which I
hope agreement could be reached...

For a far more ambitious goal, look al psycopg2 -- YUM... adaptation
roolz!!! (My secret dream is to have psycopg2 and PySQLite 2.0 support
the same API -- at which point I will care little what the DBAPI does,
since, unless a client twists my arm _and_ dangles bags of gold in front
of me, PostgreSQL and SQLite are the only two RDBMSs I really
need...!-). Considering the nationalities of the psycopg[2] and
PySQLite developers, might I suggest rekindling the goot olt Rome-Berlin
Axis, or would that automatically invoke Goodwin's Law & kill the
thread?-)


Alex

 

[Stephen Waterbury]

... (My secret dream is to have psycopg2 and PySQLite 2.0 support
the same API -- at which point I will care little what the DBAPI does,
since, unless a client twists my arm _and_ dangles bags of gold in front
of me, PostgreSQL and SQLite are the only two RDBMSs I really
need...!-). ...

I wholeheartedly second that -- my favorites, too!