It’s time for Woodie’s Framework Roundup! (It’s a Toy Story 2 reference)

p. This was a roller-coaster week for the framework mailing list. Not “Scream Machine” roller coaster but certainly not “Runaway Mine Train” either. Somewhere in the middle we have the highs and lows of this week in the framework mailing list. Let’s take a look.

* The “Proposal: Zend_Controller_RewriteRouter”: discussion continued on after last week’s roundup when Martel Valgoerad announced that he updated the classes in his proposal. Then, in a move nobody saw coming, he updated them *again* 5 days later. How does he keep up that whirl-wind pace?

* Inash Zubair started a thread “Zend::VERSION”: on a VERSION constant in the main Zend class. Matthew Ratzloff and Kegin McArthur both chimed in and discussed some of the intricacies of assigning a version number to the framework itself and exactly what that means to developers. At the end of the day, everybody thought it’s a good idea. Inash had the last word with a non-proposal on exactly how this could work.

* Ralf Eggert continued the “Routing process for multi language websites”: discussion from last week. The end result was that while everybody things this is an issue that should be dealt with, we are now not exactly sure how. I’m sure we’ll see more on this in the weeks to come.

* Benjamin Meynell started the thread “status of ZActiveRecord”: by posing the question “I’m wondering if anyone knows the status of ZActiveRecord as described in Andi’s “webcast”:” Andi responded that an OO->RDBMS was on the roadmap and of high importance to the framework team. The was some comparison to a lesser known framework (something like slides…the name escapes me) and a technical teaser on what could be coming.

* “Zend_Db_Adapter_Pdo_Mysql – incorrect limit for MySQL 3.X”: teaches us all a valuable lesson. Before posting to the list, make sure and check SVN. Chances are the bug you think you’ve found has already been squashed. (I’m telling you these framework guys are fast)

* Andries Seutens pointed out in “Manual layout”: that the look and feel of the manual was recently updated and it looks great. Andi gave props to Mike for all his hard work on this. I’m sure we all agree that it looks great and appreciate Mike’s work on it.

* Nick Thornley, in “loading default classes from Zend_Controller_Front”: pointed out that Zend/Controller/Class.php was loading classes in an obscure and non-standard way. Turns out it was a hold-over from old filename scheme. Mike quickly fixed it and checked it into SVN. Thanks to Nick for his diligence and to Mike for his quick response.

* In “Zend_Http_Client wont connect to IP-addresses?”: Johannes Schill raised a question as to why…well, read the thread title… it’s kind of obvious. Short answer, because RFC2396 says it’s supposed to be that way. That didn’t stop an interesting discussion on the subject though. Johannes proposed code to allow for the default behaviour but allow for IP addresses to be used if a flag was set.

* Of course the high point of the week had to be Andi’s announcement “Zend Framework License”: Rob Allen and Stefan Koopmanschap both voiced their approval for this move. While it didn’t generate much discussion on the list, my informal poll of the community at large seems to indicate that this announcement was met with much rejoicing.

* Ralf Eggert, in a bit of shameless promotion, announced his DevZone article on “Using Propel with the Zend Framework”: Great article Ralf, keep em’ coming.

* Duane Day started one of the larger threads of the week when he posted “InputFilter, MVC, and error messages”: It seems that Duane has questions on exactly how forms are supposed to work within the Zend Framework. There was a lot of interesting discussion on what directions could be taken by the framework to solve this code issue. (Core if you are building web applications) Christopher Thompson posted a link to code that he has written to solve this problem.

* I know it comes as no surprise to that “Zend Search Lucene”: made an appearance again this week. Herve’ Piedvache questioned it’s speed on larger indexes. He bench-marked it (informally) against an existing PostgreSQL application he has written. Apparently Lucene did not fare well. Markus Fischer chimed in with his anecdotal evidence that could point people in the right direction. Then Alexander Veremyev (whom I believe is the lead on “Zend Search Lucene” schooled us all in some of the details of Lucene. Basically, Lucene is designed to compete against the Full Text searches of database engines, not indexed searches.

bq. It’s more correct to compare Zend_Search_Lucene with Java Lucene, IBM Text Extender, MS SQL Full-Text Search subsystem, Oracle Text and so on. For some queries SQL requests are always faster, than full-text search requests.

* Ingo Reinhart opened a thread “Zend Db Adapter->insert and DB-functions or numbers”: in which he reported a problem with using now() in a query. The general consensus was don’t do that.

* Nik Wakelin tossed a potential e-grenade this week when he asked the question “Zend Framework – ready for production?”: Again with the comparison to that other framework he asked if ZF was read for prime time. The general consensus was not yet but Mathew Ratzloff said it best with

bq. Rails has a prescribed way of doing things, while Zend Framework has more of a “PHP philosophy”: the framework doesn’t enforce one way over another.

* Finally, our proposal of the week came from Alexandre Riveir when he started the thread “proposal Zend_Timestamp, Zend_Date and Zend_Time”: He proposed a “standard object type for type timestamp,date and time, for works date, time and convert sql for object types”. It sparked a short discussion sprinkled libera;;y with “do we need this?” posts. Later on, he even proposed “code”:

And that’s it. The traffic level took a nose-dive late in the week. I’m postulating that it was because everybody was at “php|tek”: but hey, could be that everything had been said.

Until next week,