Good old PHPEclipse throws me several errors on uninitialized stdClass objects, and on two arrays that get neither initialized nor are assigned a value. This is merely codestyle, but it doesn't hurt.

CommentFileSizeAuthor
initialize_stdClass.patch15.62 KBpancho

Comments

sdboyer’s picture

Thanks for this...yeah, we're trying to go fully E_ALL compliant, at least at some point. Thanks for the note.

Believe it or not, I am gonna catch up on the awesome slew of patches you've been putting in - I've just had a horribly unproductive couple days (just upgraded my OS, and everything's misbehaving).

pancho’s picture

Hey Sam! Surely I believe that... :)

I was just hoping that I don't bother you too much with my minor and medium issues, while you are implementing rather complex stuff.
I do know that I'm kind of creeping around the edges, but to do any major reworkings or implementations I still don't know enough about your plans and decisions.
What would it help if I reworked something only to find out that it is not gonna happen in Panels2. The new code wouldn't help too much for Panels3, either. So I keep on following your trails, cleaning some stuff up, rethinking some usability aspects etc. - just some obvious tasks.

Wait - there is one thing I thought about that'd be cool and doable within the Panels2 API: While we can't change the Panels API, we can completely rewrite the Panels implementations, am I right? So how about rewriting panels_page and panels_mini code in OOP? They are duplicating so much code, and would benefit a lot from an abstract class that they would implement and extend. That'd be a nice step forward, while not touching the main Panels API. I'd happily undertake responsibility for that.

Also, please inform me as soon as major changes are completed, so I can resume porting to D6 - at the moment most porting work obviously makes no sense, as the code base is such a moving target.

sdboyer’s picture

Most definitely, I'll keep you up to date. We're getting (very) close to that point, to be honest - the major rewrite of the add/edit code that I did last weekend was pretty much the last really notable thing I had on my plate. The only things remaining that do constitute potentially notable shifts in the core code are, I believe, changing it so that styles can be applied on a per-pane, as well as a per-panel basis, and possibly also the shifts to panels_page that fix the bugs with the proper numbering scheme applied to new panes when a type-specific display is being delinked from the default.

Man I have to learn how to talk about this stuff in English before I even THINK about presenting on it :)

As for starting the OO rewrite on some of the client modules - ugh. Well, believe me, the LAST thing I want to do is turn away enthusiasm. I know Earl and I are both really glad to have you "creeping around the edges," as you describe it, and I also very much appreciate how limiting yourself to those edges sucks when you see a lot of problems that really merit fixing right in front of you. But there are two things that make me want to wait: first, and more of a detail, is that I believe merlin's working on a pretty big rewrite of panels_page even right now. More importantly, though, is that I think starting on something like the full OO rewrite now for any of the included panels client modules would just end up delaying 6. At this point, my goal is to keep the changes we make to a minimum, the really nonoptional ones - precisely because I really DO want to Panels3, but I can't justify working on it until the D6 port is done and there's a stable version of Panels for everyone to hug and love in the meantime.

But, by all means, start floating ideas on the OO work :)

merlinofchaos’s picture

Rewrite: Absolutely Yes. However. After release.

I beta'd this thing on New Year's Eve. There's so much to do, but the world deserves Panels 2.0 and Panels 2 for D6. So while yes, I absolutely positively agree that rewriting the modules using OO would be a great benefit, the biggest benefit is going to be to get the absolute must-have features in so we can release and give people a module that they can be happy with. Then we can branch for D6, work there in earnest, and continue with bug fixes for the D5 version.

Offtopic from this issue: And while I"ve been absent a lot lately on this project, Pancho (sorry, burning the candle on both ends with Views), I am *very* glad to have you aboard and helping out. If you use IRC sdboyer and I both live in #drupal-views on freenode (which doubles as #drupal-panels too) which can also be used as a way to help coordinate, along with the wiki.

pancho’s picture

Oh, of course finishing Panels 2 for D5 and porting it to D6 has absolute priority. I was just looking around for some things to do that don't interfere with Sam's work. But I'm sure I'll find some other things to do, the issue queue is large enough.

I'm absolutely amazed by Panels2, and am sure that - once finished, stable and a little polished - it will be more than ever a "Big 5" module.

I'm gonna be away for the weekend, but will talk to you on IRC next week!

sdboyer’s picture

Status: Needs review » Fixed

Aaaaaand, back to the original topic issue...committed!

Anonymous’s picture

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for two weeks with no activity.