I stumbled across an interesting promotional video for eZ Publish in Western Europe the other day, hosted at Youtube it gives brief insight into how and why the system is being deployed by major organisations throughout Western Europe. It is definitely something that will be useful as a sales tool for partners in any country.
Over the coming weeks I will publish a number of short tutorials and information pieces to provide assistance to new community members who are looking for more information or need a plain English explanation of how to undertake certain tasks or projects with eZ Publish. Time permitting I will try to make an examples real-world examples rather than abstract ones.
Hopefully over the coming months I will be able to tidy these posts up further and publish them in a cohesive order over at the Quiqcorp site in our new documentation section which should be launching over the Christmas period.
I have a number of topics already on my list but I am happy to take suggestions from anyone else looking for specific tutorials.
As sites become more complex and segmented we continually see new ways to approach the design and layout of these sites. Quite often you see designers resorting to additional divs, classes and ids to try and identify areas within the site, which in many cases are semantically the same as other areas… but how do we style them?
One of the simplest and easiest ways to preserve and simplify your templates, classes and ids is to identify your subtree and node id in your body tag. Setting these classes allows your CSS editors to quickly identify and write custom selectors for specific areas of your site whilst not needing to modify multiple templates.
An example of doing so would be:
<body class="node_id_899 subtree_node_id_167">
The above can be generated from pagelayout.tpl providing your CSS editors a quick and easy method for identifying their current location and can style accordingly.
Bruce has a great post over on his Stuff and Nonsense blog about eZ Publish not displaying database connection errors but instead responding with an access denied message.
I think I would prefer to see eZ Publish resort back to an “Oops” or “Our site is under maintenance” display in the case of failed database connections, in preference to seeing either a no database error or access denied, both of which have negative impacts on visitors.
As a point of interest the admin design does display Error 50: No Database Connection on the login screen when a database connection is not present.
It’s been almost 6 months since I thought of writing an eZ Publish extension for the Scribd iPaper service but travel and client work has kept my hands tied. Last weekend I found a few spare hours to start working on a new extension called qcScribd which will provide integration into the Scribd API, support for workflow processing of new documents, delayed publishing until documents are displayable, a new datatype for handling licensing, advertising and private settings as well as javascript embedding of the iPaper throughout your eZ Publish site.
The 1.0 release is almost complete and over the following few days I will ready the code and release it back to the community in its new home at http://projects.ez.no/qcscribd
I’m not sure if it’s just me or not… but I’m pretty sure that the flash player on use throughout the eZ Publish and eZ Components sites has terrible memory leak problems.
If I load up one of those videos and watch it immediately I usually escape too much harm, but if you happen to load it up and pause it for any period of time check on your CPU usage.
I noticed this when my the temperature on my laptop raised to 72 degrees celcius and Safari became slower than sloth.
Has anyone else experienced this issue?
Seems I let my blog posts slip again, it is much the same as a New Year’s resolution you make and know you will never keep.
I have been keeping busy over the past few weeks and will put together some posts on my more recent eZ Publish thoughts over the coming days. This will include some new deployment methods we are using at Quiqcorp for creating new development environments on the fly.
I also note that Kristof Coomans is releasing a new eZ Publish 4 version of his Akismet extension based on the Services_Akismet PEAR module which will be great to see, that’s one thing I can cross of my list if someone else is implementing it 
Yesterday I was pleased to receive an email from the eZ Partner Team informing us we had reached Gold Partner status which is a great achievement for our team and something we couldn’t have achieved without the dedication and passion of our staff.
Over the coming 6 months we look forward to actively pursuing new eZ Publish opportunities, providing improved and comprehensive training materials for the greater eZ community, releasing some of our developments back to the community and persuing that Platinum partner status.
In my work and travel I have found an increasing need for VPN access to remote networks, including our offices, datacentres and client sites. In most cases these are over Cisco VPN which simplifies the process and makes it a breeze to connect new users.
Over the past while I have been trying out some other VPN products including the open source OpenVPN system for remote network access. Unfortunately the Cisco VPN client does not play with OpenVPN which requires another 3rd party application.
Until recently I have been using Tunnelblick, a small application built for OpenVPN connections which has great support for Tiger but many niggly bugs in Leopard. Recently I saw an article reviewing a new application called Viscosity, an OpenVPN Client for Leopard.
All I can say is wow. Viscosity whilst still in beta is an easy-to-use VPN client with automatic import for existing Tunnelblick connections. It connects to growl to provide notifications of VPN connection and disconnection and lives away at the top of my screen in silence when I don’t need it.
If you are working on Leopard and looking for an easy to use VPN solution I recommend looking into OpenVPN with Viscosity as your client-side application. Download a free beta copy from their site and give it whirl.
By default in new installations of eZ Publish using the ezwebin or similar packages you will have access to the blog, blog post and comment content classes – there are everything you need to get started with your own blog on eZ Publish.
As I built out this site one of the things I noticed with the blog_post.tpl was that comments were being displayed newest first, in the same manner they are on the eZ Publish website, this seems a little counter-intuitive to me and certainly breaks the design pattern established by most other blogging tools available.
To flip the order of your comments to chronological order, oldest first you can approach this in two different ways. You can modify the fetch that pulls back the comments and sort_by the published date, which is what happens by default, you just need to change the ascending/descending option.
The alternative is to accept the default fetch if that is beyond your knowledge and modify the template using the reverse operator to flip the resulting array from your fetch.
I have posted an example below of how this could be done.
{foreach $comments|reverse() as $comment}
{node_view_gui view='line' content_node=$comment}
{/foreach}
I haven’t had a chance yet to run performance tests on each implementation although I believe that the modified fetch function will be faster than array reverse within template code. Over the coming weeks I will have a chance to confirm this.
Big thanks go out to Nicolas Pastorino for mentioning to me the other day he had registered here on the blog and was unable to post comments.
I have since resolved this permissions issue so commenting should be allowed for both anonymous and registered users. It does however make me stop and think about new ways in which I can test and build out deployment guidelines that cover issues like this.
In any normal development process you have clients testing and revising your work, which of course didn’t happen in this case until Nicolas mentioned it to me. Any suggestions as to how you handle the large number of tasks required for an eZ Publish deployment?
Update: Well it seems it didn’t take long for comment spam to show up. I will lock down comments to registered users only until I put other measures in place to look after comment spam.
eZ Press recently released a new book, eZ Publish Advanced Content Management, which was authored by Bergfrid Marie Skaara and Peter Keung.
I received my copy last week and I am looking forward to having the spare time in coming weeks to thoroughly read and review its contents. It should be interesting to see how this publication fits into the greater education of eZ community and partner members.
I have recently undertaken the management of a number of Plesk 8 based hosting machines and came across an interesting problem today when working with backups. On this machine I have several eZ Publish installations, one of which I was looking to copy to a local drive for testing purposes. FTP just seemed like a waste of time… perhaps I was wrong.
I logged into the administration panel of Plesk to find that the backup utilities were not installed by default, thankfully Plesk makes that operation rather easy, but I am still at a loss as to which genius thought backup wouldn’t be an important enough feature to install. Granted most serious hosts are using 3rd party backup solutions like R1Soft, but still…
Having resolved that issue I proceeded to backup the domain in question and download the resulting file… which was hidden away in the middle of nowhere inside a filesystem that looks more like a rabbit warren than an organised, enterprise hosting platform. Upon downloading the file I reached the next problem… what to do with it..?
It seems that Plesk/SWSoft/Parallels have decided tar.gz is too mainstream for their product, so they opted for something more exciting, something without an extension. I haven’t had a time to even investigate what this crazy system does, but suffice to say if you have StuffIt Expander you are in luck. Drag the file onto StuffIt and it decompresses it into what OS X believes to be a number of executable files. Drag the required files onto StuffIt to expand those into the files you require.
Is it just me, or does this seem stupid? Maybe Plesk got it right and they just forgot to inform us about this whiz bang thing they developed… anyone have any thoughts on this?
Nicolas Pastorino recently released the eZKeyword Autocomplete extension for eZ Publish 4 and I decided to give it a whirl.
Installation of the extension is very straightforward:
It doesn’t get much easier than that, but unfortunately it didn’t work. A quick review of the module.php showed that /ajaxbackend/autocomplete_keywords was being queried for keyword results. On loading up this url I encountered a PHP Fatal Error.
[Mon Jul 07 23:51:17 2008] [error] [client 202.0.51.250] PHP Fatal error: Call to undefined function json_encode() in /home/andrewdu/public_html/extension/ezkeyword_autocomplete/modules/ajaxbackend/autocomplete_ezkeywords.php on line 66
It appears that the script was unable to find the json_encode function. Doh! json_encode was only added to PHP 5.2, which leaves a few of us still on PHP 5.1 with an extension that doesn’t work. Of course we can soon rectify this issue by providing a fallback function incase the json_encode function is not available on your system. I patched ezkeyword_autocomplete/modules/ajaxbackend/autocomplete_ezkeywords.php with the following function.
if (!function_exists('json_encode'))
{
function json_encode($a=false)
{
if (is_null($a)) return 'null';
if ($a === false) return 'false';
if ($a === true) return 'true';
if (is_scalar($a))
{
if (is_float($a))
{
// Always use "." for floats.
return floatval(str_replace(",", ".", strval($a)));
}
if (is_string($a))
{
static $jsonReplaces = array(array("\\", "/", "\n", "\t", "\r", "\b", "\f", '"'), array('\\\\', '\\/', '\\n', '\\t', '\\r', '\\b', '\\f', '\"'));
return '"' . str_replace($jsonReplaces[0], $jsonReplaces[1], $a) . '"';
}
else
return $a;
}
$isList = true;
for ($i = 0, reset($a); $i < count($a); $i++, next($a))
{
if (key($a) !== $i)
{
$isList = false;
break;
}
}
$result = array();
if ($isList)
{
foreach ($a as $v) $result[] = json_encode($v);
return '[' . join(',', $result) . ']';
}
else
{
foreach ($a as $k => $v) $result[] = json_encode($k).':'.json_encode($v);
return '{' . join(',', $result) . '}';
}
}
}
The above takes care of our json_encode issue and the script should now work. Autocompletion of the keywords ensures that you pick appropriate existing keywords each time rather than cluttering up your keywords with many similar keywords.
In summary the extension works well and thanks should go out to Nicolas for such a great contribution.
Anglican Taonga, a magazine published by the Anglican Church in New Zealand, approached Quiqcorp back in late 2007 regarding the opportunity to web-enable their magazine. They were looking to take the next step in the evolution of communication and try to provide materials to their readership in a shorter timeframe with the ability to cover more features than possible within a usual print setting.
The result of this collaboration is Anglican Taonga, a simple but effective eZ Publish 4.0 deployment to suit the requirements of the client. In addition, they can grow into this solution as additional areas of the Church look to expand their online presence.