Matt: I totally trust Sam, and if he has time to put the stamp on it then I could fast track it.
I only am interested in investing my time on PHP5.
Matt: I totally trust Sam, and if he has time to put the stamp on it then I could fast track it.
Looks like I’m being recruited into the WordPress QA department.
While Matt puts his faith in people (like me? egads!), I put my faith in test suites that reflect the accumulated knowledge of people like Joe Gregorio and Tim Bray. But of course, Matt is free to say that he knew I would say that, and that’s exactly what he meant by saying that he totally trusts me.
As I understand it WordPress 2.3 was originally scheduled for September 20, 2007, but now appears to be unscheduled. Proceeding with the original dates as templates for the moment, the plan would require any features that are to make 2.3 be greenlighted by Matt and team by August 20th, and then the patches beaten into shape over the next 30 days (presumably front loaded as much as humanly possible).
If the schedule looks anything like that, I’m in, with one proviso. I only am interested in investing my time on PHP5. This presents an interesting dilemma. Does WordPress:
What I mean by “I’m in” is that I will install the latest Wordpress, and track to SVN. I will run it against both of the test suites mentioned above, as well as to Pete’s latest work. I will try to encourage that there be a September 2007 Interop session, and that WordPress 2.3 be represented there. In the process, I will feed back bug reports and patches to any or all of the associated codebases as I see fit. If at the end, I’m happy, then I will give my thumbs up.
I’ll get started as soon as I hear that this feature has been greenlighted.
I’m honestly surprised that there’s no healthy, hearty discussion in the comments here. :)
I know what Matt’s said about PHP5, but I think that WP is just going to end up being one of the PHP apps that goes PHP5-only in its leading-edge release. Why? Well, for folks stuck in PHP4-land, WP 2.0.x is going to be maintained until 2010, and with no new features going into that codebase, it’s totally easy to leave that as a PHP4-focused release.
Some Web app is going to be the one that pushes the PHP5 domino over. I think the PHP developers know that, and they’re forcing the hands of PHP developers. Thing is, I can’t see where WP isn’t going to end up driving into that hill sooner rather than later, given the commitment to 2.0.x and the PHP4 end-of-life. I’d argue that it’s better sooner, rather than later, to make the leap and force hosts' hands.
Before you say, “Easy for you to say, Mr. User,” I’m a WordPress host myself, with about 100 installs. Do I run PHP4 on my box? Yes, yes I do. I even recently nuked-and-paved the box, and I stayed in PHP4 land. But man, if I knew that WP 2.3 [or 3.0, or whatever versioning scheme they’d use] was going to be PHP5-only, I’d have already flipped the switch.
I’m honestly surprised that there’s no healthy, hearty discussion in the comments here.
My experience is that, whatever the context, statements that you people (for some value of “you") should/must do something leads to a vigorous (sometimes healthy, often times not) discussion. If that type of discussion is your thing, I know of a few blog sites that specialize in that kind of thing.
But if you turn it around, and say that *I* only intend to do "X”, anybody want to step forward to fill the gap? Well, my experience is that such discussions are often met by the sound of crickets.
anybody want to step forward to fill the gap?
I’d be happy to help move WordPress’s Atompub functionality forward. But it seems job one is getting access to a committer. Attaching patches to a Trac ticket doesn’t seem to be the best way to get them into the 2.3 branch.
I take Matt at his word when he says that if I were to stand behind the patch that he would fast track it.
Meanwhile, it appears that my offer has been Warnocked.
Sam, as I said in my blog, we’d love to have you direct input on these issues. Things need to be fairly final before the 20th, as you said. As for PHP4 vs 5, I’m fine with making the Atom features 5-only, but usually it’s not hard to make something 4-compatible and doing so would certainly improve adoption, which I assume is the shared goal here.
When you say “I only am interested in investing my time on PHP5.” does that mean you plan to do something to deliberately break our current PHP4 compatibility, or that there’s some functionality there that you must have to make any changes, or simply on principal that you don’t plan to test against 4? You don’t have to answer here, email is fine for the conversation. We’ll figure out how to make it work regardless.
Geof, WordPress is fully PHP5-compatible, and we have been since PHP5 was first released. There’s nothing we do that should prevent you from upgrading.
Finally, if it’s worth anything, when Dougal gave us the code that shipped with 2.2, it passed every test on Tim Bray’s APE. Whatever broke between now and then we plan to fix, which to my knowledge is different than the approach by Blogger/GData, Nokia, and Six Apart. If we had not implemented anything, like our open source brethren in Drupal, Habari, Serendipity, or b2evolution perhaps we could have escaped criticism. However to keep Mark commenting I shall find a new “transition” to rail against, maybe that GPL thing that’s been getting all the hype lately.
Sounds like we have a plan. I’ll start with ticket 4191.
As a part of my testing (which will include APE, by the way), I will do a pass with PHP4. Anything that is easier to fix than to document will be fixed. Anything that is easier to document than to fix, will be documented. Anybody who wishes to pick off any PHP4 specific ticket that may be generated by this process will be welcome to do so.
A good indication of WordPress' incentive to move hosts and users of the application to PHP5 would be to sign up for Go PHP5 and publicly announce that “PHP5 is recommended” for any and all WordPress 2.2+ installations (for instance). Keeping the Atompub implementation PHP5 only would also increase the incentive and motivate hosts and users even more.
Sure, it’s possible to get everything to run on PHP4 with enough hacking and patching, but it surely can’t be WordPress' official goal to support PHP4 forever? The break has to come some day, in some version. Why wait? An optional feature like Atompub would be the ideal way to introduce PHP5-only and thus force hosts and users who wants to enable it to migrate to PHP5. Since the current codebase already runs on PHP5 (I do so myself), I don’t see the problem.
More to the topic, I’d like to see Sam getting commit access to SVN. That would underscore and prove Matt’s trust for Sam as well has his willingness and dedication to have full Atom and Atompub support in WordPress. The track record isn’t very convincing and I’d like to be convinced.
The updated AtomPub code is in Beta 1 of WordPress 2.3:
http://wordpress.org/development/2007/08/23-beta-1/
so if you’ve been waiting/hoping for this, here it is.