- Development - lists.drupal.org

Drupal SU module anyone?
by David Metzler 12 May '09

12 May '09

I had another idea for a simple module that I'm considering developing. I often get overwhelmed with drupal administration, and have many sites where admin is shared. Every time I add a module in drupal I tediously enable all access permissions for my "Admin" role. I began to think about the Linux sudo model, and thought this might be cool for dupal. Users with "Become Admin" privilges could switch to the "admin" user only. This way I could largely see the view that my content contributors do :). Does this seem like a good thing? Anyone know about duplication? I'm aware of Devel, and Masquerade but I didn't feel the need for that on the production sites, and I really don't want to expose the "Become Any User" functionality (but maybe I'm being silly here).

9 11

Strange error on drupal_set_message calls in hook enable
by Daniel Caillibaud 11 May '09

11 May '09

Hi, In hook enable (mymodule.install file), drupal_set_message throws a fatal error, but only if called in a if { bloc }. Strange isn't it ? This is the code, tested with php 5.2.6 define('DEBUG', TRUE); function mytestmodule_enable() { // this line throws Fatal error: Call to undefined function drupal_set_message() // if (function_exists('drupal_set_message')) { drupal_set_message("fct mytestmodule_enable"); } // but this one don't (if previous line disabled) drupal_set_message("fct mytestmodule_enable"); // and this one also throws the fatal error //if (DEBUG) { drupal_set_message("fct mytestmodule_enable"); } } -- Daniel

1 0

Re: [development] Issue queue links from project pages?
by As If Productions 10 May '09

10 May '09

At 08:28 AM 5/7/2009, Margie wrote: >Most people already have all sorts of anxiety about "I'm not as good as >that." I think the maintainers list could tend to reinforce the star power >of a few people, but that's a whole psychological question. I agree with Margie on this. We recently had a discussion about warming new people into the fold, trying to lower barriers to various types of contribution, and I think this structure will tend to work against that. Yes, it's largely a psychological question, meaning "interface design issue". Meanwhile, if you want to talk about wasted space, listing the "Latest 5 Issues" is useless to anyone except the module maintainers. LVX TF --- As If Productions http://www.asifproductions.com Interactive Worlds and Immersive Obsessions

6 8

Display a custom menu by code in hook install
by Daniel Caillibaud 09 May '09

09 May '09

Hi, In a module of my own, I build a custom menu, but I don't see how I could enabled it in my install hook. Is it possible ? My code is, in hook install : db_query("INSERT INTO {menu_custom} (menu_name, title, description) VALUES ('%s', '%s', '%s')", 'mymenu', t('Test menu'), t("My own menu tree for testing purpose.")); in hook menu : [...] $items['mymenu1'] = array ( 'title' => t('Entrée root 1'), 'page callback' => 'mymenu_test', 'page arguments' => array (0), 'type' => MENU_NORMAL_ITEM, 'menu_name' => 'mymenu', 'expanded' => TRUE, 'access callback' => 'user_access', 'access arguments' => array('access content'), ); [...] With this code, I have my custom menu, but I still need to activate it manually in admin/build/block. I can't call db_query on blocks table in my hook install because block doesn't exist yet. Is there a clean solution ? (creating a block entry with sql query in hook install doesn't sounds very nice). Should I have to replace all my items in hook menu and my custom menu by menu_link_save calls (under navigation menu tree) in hook install? Thanks -- Daniel

1 1

Issue queue links from project pages?
by Margie Roswell 09 May '09

09 May '09

What happened to the issue queue links from project pages? Can they be restored? Best, Margie -- http://www.BaltimoreUrbanAg.org http://www.FriendlyCoffeehouse.org http://www.packtpub.com/drupal-5-views-recipes/book

11 24

Druplicon IRC bot needs yer channels! Help stress test new code!
by Morbus Iff 08 May '09

08 May '09

As some of you know, Druplicon has been "full" for awhile, as I've denied any new channel joins due to the bot having reached a threshold that caused it to choke on handshakes. Some of you may have even recalled this factoid: [13:06] <Druplicon> When I first start, I'm in so many channels that it takes me a long time to cache all the necessary user info. This is not a fault of my code, but rather that I'm in 20+ channels. This handshaking causes me to be initially unresponsive for 2+ minutes and, if you directly address me, I may flood. I have just applied code to hopefully fix this problem. The inline comments should explain the change suitably: When the bot has a number of channels to join (nearing 20+), the initial handshake, where it retrieves information about every user in every channel, can cause it to overload and disconnect. To solve this, we join channels every 15 seconds instead, which gives some breathing room for the nick caching to take place. A nice side-effect of this is that if the bot errors out of a channel, or a new channel is added to the list, it'll automatically (re)connect without needing a restart. This now means that for every restart, Druplicon will take a longer time to join every channel (15 seconds * NUMBER_OF_CHANNELS), but it should be able to join far more. And, one of the primary reasons for restarting Druplicon in the first place was to have it join new channels - this restart should no longer be necessary due to the above changes (due to the variables cache being cleared every five minutes, which would cause the new code to detect newly-added channels and autojoin them). So, what I need is: * All those people I said "No Druplicon for you!" to ask me again. I'd like to stress test things by adding him to every channel requested and see where the next major chokepoint will be. [Note: this is a problem unique to Druplicon - I don't know of many IRC bots who are in more than 20 channels, much less those who are requested to be in, easily, 30 or more. That's how widespread our channels are.] -- Morbus Iff ( is this a cut out bath-poster Morbus, or what? ) Technical: http://www.oreillynet.com/pub/au/779 Enjoy: http://www.disobey.com/ and http://www.videounderbelly.com/ aim: akaMorbus / skype: morbusiff / icq: 2927491 / jabber.org: morbus

2 2

Menu Node API
by Ken Rickard 08 May '09

08 May '09

We had a client who wanted the site's menu system to be used as the sole organizing principle. To allow editorial access based on menu items, however, we had to figure out how to write JOINs from the {menu_links} table to the {node} table, which is currently not supported. One result is the Menu Node API [1], which maintains a {menu_node} table that can be used by modules (such as Views) to write JOIN queries based on menu relationships. My question to the list: Should this tracking feature by in core? Please respond here: http://drupal.org/node/457450 -- Ken Rickard agentrickard(a)gmail.com http://ken.therickards.com [1] -- http://drupal.org/project/menu_node

1 0

Re: [development] Menu Tree Access Control?
by Simon Roberts 08 May '09

08 May '09

> Date: Fri, 3 Apr 2009 00:11:51 -0400 > From: Andrew Berry <andrewberry(a)sentex.net> > Subject: Re: [development] Menu Tree Access Control? > I'm working on a site nearing completion which uses Books and > book_access for a case similar to yours, though without quite as much > granularity. We've been quite happy with book_access. If you pursue a > more fine-tuned system for permissions within a book outline, let's > get a group going and see what we can come up with, as I think it's > something my clients would be interested in. I've also been doing fine-grained book-page access, per-user and per-role, and ended up coding more than I would have liked. (Writing ACLs on the fly, and twiddling with menu-entries during theming - uugh). No idea if this is useful, but I saw this new project pop up a few days ago http://drupal.org/project/book_page_access

4 5

global object to push stuff into <head>
by Ivan Sergio Borgonovo 07 May '09

07 May '09

I was looking at nodewords (aka Meta tags) and I developed my own mini module to push stuff into <head>. nodewords grab informations from "core" or very popular modules to add stuff into meta and it provides a basic api to other modules. But somehow meta are a standard part of a web page and drupal provide just drupal add_html_head/drupal_get_html_head. So there is no way to edit in a structured way head. A common use would be to add keywords or replace them and support keywords/description for content that is not node-based. If there was such a structured object to deal with META it would be easier for modules and non-node content to place stuff there. And in spite of writing an "additional" snippet of code to nodewords to support different modules, other modules will just have to pass what they want to this object. I already wrote a module that provide such object to give a structured access to HEAD and provide some helper to easily access, overwrite, add description, keywords and canonical url + support for nearly arbitrary head tags. I`m going to build up a nodewords clone to experiment with an hybrid approach that will offer modules to push stuff into HEAD and pull stuff from core object as nodewords already does. Of course I already have some other module that push stuff into meta. There could be a head_alter hook or something similar. I was wondering if a more structured access to HEAD would gain any traction in core or there is already someone working on it to get an idea of the direction core may take. thanks -- Ivan Sergio Borgonovo http://www.webthatworks.it

3 5

Proposed Apocrypha
by Jeff Eaton 07 May '09

07 May '09

Crell's Law: If an API must be use-case-optimized, make it swappable & tailor the default for cheap shared hosting. High-end sites can swap. Eaton's Corollary: If an API is swappable, write two implementations. APIs with one test case are rarely flexible enough for the second. --eaton

11 13