Re: [PDO] PDO2 as a set of PECL extensions, rather than PHP core?
[email protected] (Ben Ramsey)
| Newsgroups | php.pdo |
|---|---|
| Message-ID | <[email protected]> |
On 1/28/08 12:54 PM, Dan Scott wrote: > On 28/01/2008, Steph Fox <[email protected]> wrote: >> Hi Dan, >> >>> Those non-CLA signing people could even change the code in core, too, >>> if they felt that it was warranted. They would be contributing to a >>> derivative work, not to PDO2 itself. >> That's where the plan falls down. It would be **symlinked**. > > But it doesn't have to be symlinked. That's just the way that things > are done currently, right? Code could just be continually copied and > merged from the PECL repository into core to maintain the core PDO's > derived state. > > Alternately, PDO could be pulled out of core. When you download and > install Perl, you don't get DBI - you have to get that separately > (well, maybe not on Windows, depending on the distributor's > package...). I'm pretty sure the same is true for Python (as there is > no core DBI package in Python, just db drivers that implement the DBI > interface). Does it really matter if PDO is part of the php-src CVS > directory or not? > What if PDO2 was essentially an API for writing a DB driver (a la pecl/pdo_user, though not providing the API at the userland level)? I know it somewhat serves that function now, but PDO2 would be a better API with better documentation. The PDO2 code would be in the core and not under any kind of CLA. Then, the various database vendors could develop drivers for PDO2 based on that API, and the code could live on their sites or in PECL, depending on how they choose to license their code. It would benefit them more, however, and allow them to better compete against other database companies if they choose to open source their PDO2 drivers and provide them to the community from PECL, so I think there is a big incentive for them to do so. Then again, maybe I'm over-simplifying things. :-) -- Ben Ramsey http://benramsey.com/