Menu columns

Updating status.

subsc

I doubt this will be approved while it requires a core edit. Is there any other way to implement it?

I just implemented this same functionality on a client's site in D7 using a custom block. Would it be possible to implement like http://drupal.org/project/menu_block ?

Hmm. I didn't realize you were changing the menus themselves. What is the purpose of doing that vs. changing the menu display during theming?

The method I used to break the Main Menu into columns and render in a block:

/*
 * Create Main Menu block content
 *
 * Gets the first level of the main menu, breaks it into multiple lists, two
 * items per list, and renders all lists.
 */
function site_blocks_main_menu() {
  $tree = menu_tree_page_data('main-menu', 1);
  $content = '';
  while(count($tree)) {
    $menu = array_slice($tree, 0, 2);
    $output = menu_tree_output($menu);
    $content .= render($output);
    $tree = array_slice($tree, 2);
  }

  return $content;
}

I suggest/ask because I'd like to see this work w/o core edits.

If the hook is needed in core, then file a patch for core.

In my opinion, I don't think we can approve project applications for modules that do not work with core 'out of the box', under the existing process. End users expect to be able to download and install without having to do any additional coding.

The existence of other modules with required core patches is unfortunate ... but I'd bet money that not one of those modules came through the project application process. Once a user has one module approved, they are granted the right to publish other modules without review. If they then choose to publish modules that require core patches, then that is their perogative ... but it is not something that is widely encouraged (or even supported) within the community.

From my perspective, if you can make things work without the core edit, then the application can be considered ... if not, I'd suggest applying with a different project.

In the meantime, I'll add this scenario to the items for discussion in the Code Review Group, and ask for a few extra opinions.

EDIT: Consider myself enlightened. There are mixed opinions on the core hack issue within the code-review group.

If the functionality can be accomplished without a core edit, that would be preferable ... in this case, be sure to continue chasing a core patch if it is a hook that others would find useful; but it sounds like we wouldn't make the core edit a blocking issue. the core edit being a blocking issue is up for debate within the Code Review Group.

I agree with jthorson, the core hack requirement needs to be removed before this can move forward. We have an established process for editing core through the Drupal core issue queue. Part of that process is review by the security team. We also have an established process for sharing more experimental code like this, and that's what sandboxes are for. If people are technical enough to use Git to get your code and want to follow your advice to alter Drupal core, that's their choice. But if you start encouraging less technical users to do that, you'll be creating a huge headache for the security team.

On a quick scan of what you're doing in that core hack, I believe you can do the same thing in D7 without a core hack, since D7 calls drupal_static() in menu_tree(). I suggest moving your code to D7 as a way to remove the need for a core hack and move this forward.