PDO2 as a set of PECL extensions, rather than PHP core?
[email protected] ("Dan Scott")
| Newsgroups | php.pdo |
|---|---|
| Message-ID | <[email protected]> |
Thus far the response to the proposal of a CLA for the spec and code for PDO2 has not been particularly positive. I believe the idea of PDO2 itself, however, would be very beneficial for the PHP community. I would therefore like to put forward another proposal that might allow PDO2 to move ahead. Of course, it doesn't have to be considered now, if others feel that there is still merit to continuing the CLA debate. Given that some other languages (for example, Perl and Python) have successfully developed standardized database interfaces outside of the language core, would it be possible to consider an alternate approach for PDO2 - develop it outside of the php.net core repository purely as a PECL extension? (The Python DBI lives a bit of a double life, really - the spec itself is a core PEP but the drivers are all extensions, if I understand it correctly). While this means that a person who compiles PHP themselves will have to go to extra lengths to compile PDO2 + the PDO2 drivers they require, in practice, many PHP developers and deployers rely on their distribution's precompiled packages and the extra step would simply become "apt get install php-pdo2-foodb" (or the like). Given that some of the commercial database servers already face installation hurdles due to their choice to distribute binary client libraries, this is probably not that big of a hurdle in comparison to finding and installing the right binary client libraries. As a set of PECL extensions, PDO2 and its drivers would be documented in the PHP manual and still be visible on the PHP site; bugs could still be filed through the normal PHP bug process; the unit tests could still be run and gcov charts could be built to show code coverage process; the code could still be stored in the same PHP CVS repository. The only difference is that PDO2 and its drivers would be extensions. If there are technical or procedural reasons that this proposal would be impossible or unacceptable, I apologize for not anticipating them. -- Dan Scott Laurentian University