Remove all themes but Garland and Stark from core

To be clear, you don't need another theme engine to use Chameleon and Marvin. But, other than that nit, I agree 100%. Hanging out in support as much as I do, I see tons of confused newbies who don't realize that the prevalant PHPTemplate snippets don't work on Chameleon. Keeping core to just PHPTemplate themes would help that a lot.

Michelle

Ah, sorry about that wild assumption without checking. But yeah, avoiding that confusion is a real win here.

I vote for removing all templates but Garland and Minelli from core. They haven't been improved since Drupal 4 AFAIK (some if not all still use tables for layout purposes) and they look like themes from before the year 2000. Keeping them in core only makes core larger than necessary.

I'm with Xano here. Replace them all. Pushbutton is phptemplate based but it's using tables and Bluemarine is just dated. Add in a few starter themes that are easy to work with and we should be fine.

+1 on this.

yeah, I'd agree, all themes except Garland/Minelli are outdated and are doing us a disservice now. Just wanted to start small and get some stuff removed before we even have alternatives ready.

Agreed - +1!

How did I miss this one.

So. Who's offering to maintain these themes in contrib for the people who are still using them in production?

Haha. As if…
But seriously, who is maintaining PHP3 at the moment for the people who are still using it in production?
Will you really require these to be kept alive?

We don't provide backwards compatibility for code, only for user data. If we're removing core modules, then in 90% of cases we need to provide an upgrade path for user data, however in the case of throttle module where it had been superceded by better options (block caching etc.), and where having it in core was worse than not (all those extra checkboxes all over the place on already confusing screens) - we just took it out. I think these themes fall into that latter category.

If there really are sites out there using them in production, then those which are using hacked up versions of these themes have work to do either way - whether re-applying changes to the 'base' theme, or following the upgrade instructions for their 'custom' theme. The number of sites actually running one of these themes in any recognisable state that I've seen in the past two years I can count on one finger, and those have plenty of better alternatives to choose from in contrib.

And they aren't maintained in core either:
http://cvs.drupal.org/viewvc.py/drupal/drupal/themes/bluemarine/
http://cvs.drupal.org/viewvc.py/drupal/drupal/themes/pushbutton/
http://cvs.drupal.org/viewvc.py/drupal/drupal/themes/chameleon/

Chameleon has moved to contrib and John Albin has volunteered to maintain it (see #374650: Chameleon broken by hook_page_alter).

http://drupal.org/project/Chameleon

I guess this means Chameleon can now be removed from core?

Patch (minus the binaries).

The last submitted patch failed testing.

[cry] Nooooo, my loved baby! [/cry]

Ok, although she had some attraction 5 years ago, she's now adult and ready live in a countryside with her equally ugly husband Marvin.

Somebody said "taken to the pasture"?

Ok, so I own http://drupal.org/project/Chameleon

I created a request in the bluemarine-in-contrib issue queue to get access to it: #397274: Project maintainership

But, the Pushbutton contrib in CVS doesn't have a project page. And from the $Id$ tags it looks like Dries owns it. So I started #397478: Change ownership of contributions/themes/pushbutton to get that sorted.

@catch: it's strange that removing the content's of those files cause the Drupal installer to stop working. Odd.

I now own http://drupal.org/project/bluemarine. I'll update the code in CVS shortly.

What about the Stark theme? I think the title should reflect that.

I think Garland may be referred to as a "Classic" Drupal theme to keep in there.

Good job guys.

sad to see bluemarine removed from core.... it was a great, simple (starter) theme, very drupalish (blue), which was just outdated and needed updating to current best practices.

we need a new starter theme like bluemarine in core (another issue...)

I don't see what was so great about Bluemarine. Garland does a far better job. Besides, we can linger in the past and decide not to remove those old themes from core, or we can remove them to prevent core from looking old and foolish for newcomers.

I now own:

http://drupal.org/project/bluemarine
http://drupal.org/project/chameleon
http://drupal.org/project/pushbutton

I'm just waiting for the packaging scripts to create the tarballs for the -dev releases I created. :-)

Patch attached. I'm betting the testbot will fail to apply the patch because it removes image files. I'll re-roll without them if it does.

The last submitted patch failed testing.

Yep. Re-rolled patch without image references. Those will have to manually deleted.

The last submitted patch failed testing.

We have contrib projects and a reliable maintainer. If the patch doesn't actually work, cvs remove will do.

Here's a patch which just updates the CHANGELOG.txt. (catch asked for this in IRC just now.)

Doh! Here's the correct patch.

As we add new themes to core, we can add them into this same line in the CHANGELOG.

"Replaced the bluemarine, chameleon and pushbutton themes with Stark" -- i do not see the logic ("with Stark" part).

@ kika - We don't mean that stark is replacing those three themes... for now, we're removing old cruft, with the plan of getting a few new themes into core. And hopefully a new default theme (that remains to be seen, though!).

Stark is an excellent theme for a person wanting to develop a new theme; it's terrible for half the people that download Drupal to quickly set up a new website though. That's why it will be installed with core, but not enabled. We hope there is a new killer theme for the default core theme.

I don't see what was so great about Bluemarine. Garland does a far better job. Besides, we can linger in the past and decide not to remove those old themes from core, or we can remove them to prevent core from looking old and foolish for newcomers.

ever tried learning theming with garland???

It's a lot harder because of the "unneeded files" (e.g. color module support) while bluemarine is just 5 .tpl.php files, 3 images (1 logo, 2 screenshots), 1 CSS file (+1 CSS file for RTL support ), LICENSE and README, and the .info file, while still supporting the full drupal toolkit.

this made it perfect for a new themer to get used to drupal, in an easy way,
also, since it had clear CSS rules, it was easy to understand.

but the best thing about bluemarine was, it had that nice, warm drupal feeling, unlike charmeleon and pushbutton.

i'm not cracking down on Garland (unlike charmeleon and pushbutton) here, each theme has it's own strengths and weaknesses

@ Alexander Pas - the problem, though, is that even having 5 .tpl.php files can be quite daunting for a first-time themer. The goal is to have more CSS-based themes, and have the whole theming idea be very designer-friendly, and lean on the PHP. Personally, I would hope we can make the core theme simply be a CSS file, maybe a few images in an 'images' folder, and a .info file. Especially since that would show off how amazingly designer-friendly the core output is.

It would also be nice to have a good README.txt file in the core default theme, explaining how people can easily extend and adapt the core theme to do whatever they'd like to do, including copying in a .tpl.php file or two.

There are some tests that check switching of themes that need to be updated as well.

Added updated block.test tests. And updated CHANGELOG wording.

Looks good to me, waiting for test bot.

@ Alexander Pas - the problem, though, is that even having 5 .tpl.php files can be quite daunting for a first-time themer. The goal is to have more CSS-based themes, and have the whole theming idea be very designer-friendly, and lean on the PHP. Personally, I would hope we can make the core theme simply be a CSS file, maybe a few images in an 'images' folder, and a .info file. Especially since that would show off how amazingly designer-friendly the core output is.

First of all, I completely agree with you. bluemarine shoudn't be a starting point for theme developpers (neither should garland), but a simple CSS+images only stylesheet should be, however, when wanting to expand the theme, the transfer from CSS+images to e.g. Garland is still way to big, bluemarine would have filled thig gap perfectly.

remember that some people like to learn from example, while others learn by reading ;)

I was thinking of a learning curve like this (included stark as the "no theme")
stark >> (new CSS only theme bluesky?) >> bluemarine >> garland

Guys, have a look at http://groups.drupal.org/node/16200 for the bigger picture on new core themes. The issue queue is not the best place to discuss "big ideas" and you'll find a lot more info, audience and context over there. Thanks!

Actually, a more updated plan is on the Community Initiatives page under Design & Theming:
http://drupal.org/community-initiatives/drupal-core/theming

And, if I can plug my DCDC session video… :-) http://dc2009.drupalcon.org/session/how-kill-blue-marine-new-core-themes...

I support removing these themes, but I'll let webchick do the commit.

I think we should move these themes to contrib so people can continue to maintain and use them.

The next step could be remove support for different template engines and standardize on PHPtemplate? Just thinking up loud.

@Dries: JohnAlbin is already taking care of them in contrib. As much as I love PHPTemplate, there are some people firmly in the smarty camp and I think you'd have an outcry unless it could somehow be made to work from contrib.

Michelle

We could use a new issue for removing other template engines in Drupal. Let the smarty folks duke it out there.

*drumroll*

I now present to you: Drupal 7, 75% less ugly than before!! Hoorayyy! :)

Great job, folks!!!

Incidentally, for posterity, here was my case for this patch. Thanks, Dries. :)

One of the groups of people I spend a lot of time talking to (both at
Drupalcon, and IRC, and "real life" and...) are designers and themers.
Drupal has quite a few pain points that prevent designer adoption, which
hurts our ability to compete with products like WordPress that have tons of
beautiful themes to try out.

One of the primary reasons for this is that designers get a face-slap of
ugly the first time they install Drupal and go looking for something other
than Garland to style their site. Pushbutton? Ugh... Table-based with
PowerPoint '97 *arrows*? Chameleon? Also table-based, AND none of the theme
snippets work for it, since it's not PHPTemplate-based. And, well.
Bluemarine just is looking pretty dated these days. :) This first impression
often leaves designers scrambling as far away from Drupal as possible, to
prettier looking products until they finally hit their limitations and
begrudgingly come slinking back.

There are a number of things that members in the design/theming community
are doing to help address these pain points (improved documentation, a
"theme drive" to help get more awesome themes in contrib to choose from,
etc.), but a *really* simple first step is simply removing all of these ugly
themes so this first impression doesn't sting quite so much:
http://drupal.org/node/315533. I think if we did this, it would send a big
message to the design community that we're taking their issues seriously,
and would encourage more patches from them. I've even coerced John Wilkins
to take over maintenance of these themes in contrib so that people currently
depending on them can still get them in D7.

What say you? :) Can I commit that sucker? :D

Go Webchick!

Michelle

Good news, glad we could do this without requiring new themes to replace these. Less is more!

My happy dance ensues! Webchick, I was going to track you down in DC to make my case for this. Looks like you were one step ahead. Thanks :)

Automatically closed -- issue fixed for 2 weeks with no activity.

FYI: This (minorly) "broke" update.module: #490562: Hard-coded list of core themes for update module is now stale (and yes, I know, #253501: Tests for update.module would be nice). ;) Just something to keep in mind when folks are adding/removing core themes. Thanks.