Hello,
First up, I have been quite outspoken against CLA'ed code in PECL in
the past. With the current proposal I am willing to reconsider this
position. The main difference to me is that is that I expect the
different vendors to take a much more active role in the development.
This means that we can expect them to fix any issues that need quick
attending (like a ZE change) to not break the build etc. Also we are
getting much more. If things do not work out for us, we can always
fork and work off of a hopefully solid base. I think we have enough to
gain to give it a try.
So one thing that I am not clear about is why we are talking about
PDO2. There are two arguments I have heard in this context:
1) none of the drivers are really compatible in the sense that they
are all full of needless differences
2) the pdo callback mechanism is not ideal for all rdbms
So number one does not really make PDO2 necessary. We could just
define a common standard now and add a "compatibility mode". This way
it would mean that current users could much easier get the benefits of
a common standard behavior without breaking too much code that depends
on the old behavior.
For number two it would take a few guys to poke all the different
peoples inside the connector development groups of the database we are
supporting to learn how to best build a callback mechanism that plays
well with all databases. Now this seems like quite a bit of work and
as long as we do not have someone lining up to take that job it does
not seem feasible to wait for PDO2. I presume that the intent of the
original proposal was that because of the CLA the developers of DB2,
Oracle, SQL Server, MySQL etc. could be allocated to this task. Maybe
Zend also planned to allocate someone? Now that we do not have a CLA
on core and spec, it seems like this plan will not pan out.
So until someone steps up, we should stick to work on PDO v1. This
means continuing the work that Wez did on specing out PDO v1, adding
documentation and tests (is this something where Oracle, IBM, MySQL AB/
Sun and Microsoft can help out?) etc. With the tests we implicitly
also define a common standard behavior. Furthermore we can work on
expanding the API for fetching meta data and other things. In the end
it means to me like all the vendors have enough to get started right
now and the community also has its work mapped out (though if someone
is willing to pick it up is another story). If some of the vendors are
motivated to jump start this process by hiring someone out of the
community is of course an option for them to consider.
regards,
Lukas
--
PHP Internals - PHP Runtime Development Mailing List
To unsubscribe, visit: http://www.php.net/unsub.php
