Admin form fields not getting rendered at bottom of node form like they should

Still valid ...

This removes the authoring information and puts the publishing options at the bottom.

Or could we remove the divs and put a class on each fieldset?

New patch that fixes a number of issues with the theme_node_form() function:

- calls form_render() on the 'authoring information' and 'publishing options' elements BEFORE running the catch-all form_render() - if it's done after (as it was previously), the call is useless and will never yield any output.
- wraps an isset() check around elements that may not always be displayed, depending on the privileges of the user. Also an empty() check before outputting the 'admin' div. This gets rid of the empty divs that were getting outputted.
- Places the admin controls at the bottom (which is what I assume the previous code was trying to do, but was failing to actually do), and the submit buttons at the very bottom.
- Puts in a bit of comments explaining that the code's purpose is to move certain fields to the bottom of the form - took me a while to figure out that this is what was trying to be achieved, it wasn't really that obvious.

This patch has Mac newlines on the changed code (and some extra), but Unix newlines on the unmodified code. This makes the patch larger (some lines have only the line ending changed) and we wouldn't want the inconsisntent newlines committed.

Newlines are now all unix-style, unnecessary lines are gone. Apart from that, same patch.

Added a whole heap of new-lines (i.e. \n) to the outputted HTML, and added a missing </div> at the end.

Why is that function called 3 times to begin with? What is the purpose of the two subsequent calls?

Uhhh.. it's not called three times, it's only called once. You mean theme_node_form(), right? As far as I can tell, it's just called by drupal_get_form(), which gets called in the node_form() function.

The original report reads: "There's a bug in theme_node_form() where all of the node's form elements are rendered in the first call to form_render(). The two subsequent calls do not output anything.". It's what made me believe it is called more than necessary.

Other than that, you're patch makes sense! Please set to RTBC (and I'll commit it) if we can ignore the comment above. Thanks.

Ahh.. yes, form_render() is getting called three times in the original code, AND in this patch. However, the "two subsequent calls [that] do not output anything" in the original code, are now (i.e. in this patch) being called first (before the third form_render() call, which is a catch-all call), and so in this patch they DO output something. Hence, the bug is fixed.

I'm pretty confident that this patch is ready to go in (just tested, still applies and works fine), so RTBC.

Committed to CVS HEAD. Needs to be backported to 4.7.0, IMO.

applied