WebCMS (OpenText Management Server)

STF Child Element Of

Thu, 05 Nov 2009 16:14:57 -0500

This little tidbit actually came as a suggestion from a colleague using OpenText at PennState University [http://www.psu.edu].

Quick preface note, when requesting that an editor enter a URL for a link, keep in mind that you do NOT have to use an anchor element for this. Standard Text Fields can be formatted to accept different values, including dates, emails, and best of all, URLs. This means the content entered by the editor will be rejected unless it contains an http:// and is void of spaces.

For my specific project, I was connecting images with corresponding URLs. For instance, I would have a picture of a professor conducting experiments and a corresponding link to a blog article explaining his latest research. There were multiple images with multiple corresponding links. The images and URLs needed to be connected in someway to ensure that the correct link was matched with the correct image.

Enter "Child element of".

I created the placeholder for the first image: img_main1
I created the corresponding link using a standard text field: stf_main1

While creating the stf_main1URL, I specified the "Type of field" as "URL". Then, on the same panel, under SmartEdit Properties -> "Child element of", I choose img_main1 from the dropdown menu.

Inside the template I placed one red dot for the img_main1:

<%img_main1%>

SmartEdit users see one red dot for the image, but when they click on the red dot, both fields (img and stf) appear:

This idea was the spring board for many more "custom" red dots that I use on the site. It's proved to be a handy grouping tool, in place of the broad "Edit Elements via Form" and the messy look of more than 5 red dots on one editable page. So far, it appears there is no limit to the number of Child Elements a parent element can contain.

Changing Standard HTM extension

Wed, 04 Nov 2009 15:56:04 -0500

By default, OpenText publishes all pages as foo.htm. Changing the extension from htm to html is a quick simple process.

Of course, you can change the extension by appending the file name on the property panel of each page:

Though this is NOT best practice, as it requires redundancy, and only changes the extension at the page level instead of the global settings. This means pages that have not been named will continue to publish as .htm, resulting in a mixed bag of page extensions on a single site. Additionally, changing the extentions in the file name can confuse OpenText's publication process so that you might end up with a double extension.

Instead, I change the extention in three places in SmartTree, just to ensure every panel is reflecting the same desired extension: html.

1. PROJECT VARIANT: Administer Project Settings -> Project -> Project Variants -> HTML (or whatever you've named your project variant) -> Edit Project Variant (Action Menu) -> Standard extension of published files = "html"

2. PUBLICATION: Administer Publication -> Project -> Edit General Settings (Action Menu) -> Standard extension = "html"

3. CONTENT CLASSES: Administer Content Classes -> navigate to your content class and expand node -> Templates -> HTML (or template name) -> Edit Properties -> Standard extension of published files = "html". This step will have to be repeated for your foundation templates only.

Have more ideas about this? Please leave a comment below.

Asynchronous Process Woes

Thu, 22 Oct 2009 16:41:43 -0500

Boy did I have a mini meltdown over this one!

One day, publication ceased on one of our projects. I contacted RedDot Support. For 8 days they had me copy the project, export the project, rebuild the publication, publish to a different target, publish target on a different project, remove workflows, rebuild workflows, add project variants, clear XML and page cache, clear out RedDot Temp files, basically an exercise which nearly drove me insane.

And then, on day 8 of no publication, I happened to notice an Asynchronous Process (SmartTree->Administer Project Structure->Project->Show Asynchronous Processes) that had been started one week prior and was "stuck" in the active phase for 8 days. The exact amount of time since the last successful publication.

Normally, with these process you can kill them by clicking on a convient X in the SmartTree.

However, this process was missing the X and there was no way to stop the stalled publication.

I sent a last final email to RedDot Support: "Please send me documentation on how to stop a stalled asynchronous process". And guess what? They had the perfect documentation on how to clear these jobs. The process involved deleting fields from the CMS database and then allowing the back logged publications to run their course. Though I have the documentation, I can not post it due to RedDot copyrights. The documentation is readily available from the Support office.

So, if your project stops publishing and you've cleared out the caches and checked the queues, try looking at the asynchronous processes, because that might be the source of your publication woes.

Check URLs

Mon, 05 Oct 2009 17:27:29 -0500

VCU has purchased a Check URLs module for RedDot SmartEdit. Check URLs will check all the links for 404 Page Not Found errors to ensure that all your links are valid.

Pages can be checked one at a time in SmartEdit by opening any RedDot, choosing "Actions" from the top menu, then "Check URLs".

A dialog box will popup, showing a list of links on the page if any exist. Click "Check URLs" to start a process that will check the URLs, sending the results in an email. Nothing will appear in the dialog box if no links appear on that page.

Want to check all the URLs on the entire site? This can be done in SmartTree.
SmartTree -> Administer Project Structure -> Project -> Actions Menu (Check URLs for Entire Project)

Generating this list may take a few seconds depending on the size of your project. A list of URLs will appear, allowing you to start an asynchronous process checking the URLs for errors, the results will be emailed by the system.

Now here is the cool part - the generated email will list all the links and next to them either "correct" or "broken". But instead of going through the entire email to get a list of broken links, log back into SmartTree and click the "Check URLs" function again.

This time the results will be grouped for you between "Broken links" and "Working Links". Click on the pencil next to the URL to change the URL of that link. All the links can be changed from this screen without having to navigate into each individual CMS page.

Publishing Reports and Errors

Tue, 22 Sep 2009 14:51:44 -0500

Page # is empty, maybe not approved?
Log into the CMS. From the left menu choose Tasks -> # Edit Pages -> Check all complete pages and click on "Submit to Workflow" button. If there is an approval flow for your department, your approver will need to release your page before it can be published.

PageId # without Content, not published
The CMS will not publish pages without at least one content element populated. Note: Anchors (anc_) and lists (lst_) are structural elements and will sometimes cause the publication to fail.

Page # not linked in HTML/ENU
During publication, the CMS follows lists and anchors when publishing a website. Checking "publish related pages" will look for any lists or anchors and publish the attached pages. If a page exists that is not connected to a list or is incorrectly attached to an anchor, the page is considered "unlinked". Unlinked pages (orphan pages) must be manually published. Unlinked pages are generally discouraged, since they are not part of the website structure and can be easily deleted.
Note: HTML/ENU is the project/language variant used for publication. HTML is the project variant checked by default during publication, though others may exist. Project variants define the format of a project's published pages.

Page not changed
If you changed an image or content on a page and try to publish but the server returns a report saying the page is already published you will need to clear the page cache. Do this by selecting "Administer Project Structure" -> Project -> Action Menu (Clear Page Cache). The system will ask "Are you sure you want to clear the page cache?" - click "Yes". The system asks this question because the operation taxes the CMS server. Wait about 10 minutes, then republish. If your pages are still not publishing, make sure your "Tasks" folder is clear and republish once more. If your changes are still not publishing contact Web Services.

Blank Publishing Report
The "images" folder, a subdirectory of one of my sites, had temporarily been set to read-only. When the CMS tried to publish any page of this website, it returned a blank publication report which was neither "successful" or "failure". I cleared the cache, and tried to republish but I received another blank report. By logging into my external server with an ftp client I realized that I did not have access to that folder. The CMS stops publication at the first error from the server. In this case, the error was access to a subdirectory, therefore nothing was published, not even to other directories. To fix this problem, remove that folder from the publication package and from the folder structure OR get permissions to the folder.

Have you seen other errors? Leave a comment or email me: tlfrance.

SmartTree Favorites

Mon, 21 Sep 2009 13:15:38 -0500

Here is a cool little tip I use frequently. SmartTree allows you the ability to set any page or content class as a favorite for quick future referencing.

In SmartTree click on a frequently used content class or page. From the "Action Menu" choose "Add to Favorites"
To retrieve your favorites list, click on Administer Project Structure -> Project -> Action Menu ("Select from Favorites")

This will popup a list of tagged pages, from here you can click on any of the favorites to be taken directly to that item in the tree. This prevents wasted time in opening embedded nodes.

Style Sheets in Firefox

Wed, 16 Sep 2009 14:39:34 -0500

In general, advise your editors to use Internet Explorer 7.0. If you have used Firefox to log in to the CMS you may have noticed that your stylesheets do not translate through the CMS. The following steps will explain how to link your stylesheets in the CMS so that they appear in Firefox when logged into webcms.vcu.edu/cms.

In your foundation template, find the code that links your stylesheet:

link media="screen" type="text/css" rel="stylesheet" href="styles.css"

Replace style.css with a container element = con_styles
Keep dynamic container and target contatiner unchecked.
After you have saved and exited your foundation template, create a new container template "stylesheet". Paste your style sheet inside this container. Click "OK" to save.
Expand the "stylesheet" content class node. Expand "Templates". Click on the template named "HTML HTML".
From the Actions menu in the right column, choose "Edit Properties".
Name = CSS
Standard extention of published files = CSS
Set template area marks in page = uncheck (checked by default)
Insert placeholder for page in container = check
To attach the new "stylesheet", open "Administer Project Structure" section, open nodes until you find "con_styles". From Actions menu choose "Create and Connect Page". Assign template of "stylesheet". Click OK. (You do not need to provide a unique name when connecting the stylesheet to the container)

Deleting Pages in the CMS/Deleting Pages on the Live Server

Wed, 02 Sep 2009 13:27:59 -0500

When a website is established, a content editor may decide they no longer need a page. The editor can delete the page in SmartEdit and republish the site. When the site is republished, the page will be removed from the navigation. The page no longer exists in the CMS project. However, the page may still exist on the end server. Additionally, that page might still be crawlable by search engines. If a user has bookmarked that page, the page will not return a 404 error, because it still exists on ther server.

OpenText CMS has a feature that will delete all the pages on the end server that have been deleted from the CMS project. This is called "clean up live server". (Note: "live server" in this case is not associated with "Live Server" - an additional product from OpenText)

WARNING!!
Do not use this feature if you use subdirectories or multiple publishing packages. When running a publishing job, "clean up live server" will delete entire directories from the live server.

To install on Server:

Delete all the files in the server folder RedDot -> ASP -> RedDotTemp (this has been done for VCU)

SmartTree setup:

In SmartTree -> Administer Project Settings -> Project -> Project Variants -> HTML -> Edit Project Variant
Under Language Variant check "clean up live server"
Republish files and check live server for deletion of old files

In my testing, "clean up live server" works fine for small, flat websites.

So what about those sites that have directories? Once a week I do the following:

SmartTree -> Administer Project Structure -> Project -> Actions Menu -> Clear Page Cache
Publish entire site from Home page, publish following and related pages.
Go into live server and look for any pages which do not have the most recent publication date. Research pages and delete.

Browser Issues (continually updated)

Mon, 10 Aug 2009 17:09:52 -0500

IE8 - use compatibility mode (08/10/09)

Firefox - all versions

no SmartTree
no access to Navigation Manager or Edit Navigation Order in SmartEdit

Firefox 3.5 - "OK" "Cancel" buttons disappear in text editor.

Strange characters on the published page

Wed, 01 Jul 2009 15:00:58 -0500

A number of editors have reported "weird characters" publishing inside their text.

The CMS text editors (RedDot, FCK, and Telerik) have character sets determined by the OpenText server charset. In most cases, WYSIWYG editors like Dreamweaver automatically declare the pages charset when the design is developed. The current, most common charset is UTF-8.

The server's charset is not UTF-8 and therefore characters like blank spaces and quote hashes come up as these weird characters.

To Solve:

1. change your meta tag in the HTML header to

<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1"/>

OR

2. In SmartTree -> Choose language variant -> "Publish charset" = UTF-8.
Header should be:

<meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>

More available on this topic and other Telerik quirks at the unofficial CMS blog:
http://www.reddotcmsblog.com/telerik-radeditor-quirks

Before creating new pages

Mon, 22 Jun 2009 15:47:35 -0500

When creating new content classes it is important to fully establish their connections before creating new pages. When you attach a reference to a placeholder, for instance, each NEW page will reference the placeholder, but the existing pages will not pick up the reference. Here is a list of actions that you should predefine before creating new pages:

Preassign content classes in the foundation
Assign Detailed Authorization Packages to lists AND Add Global Authorization Package to tree
Make sure Navigation Manager is on and your Master Pages are created before you attach the first page to the nav structure.
Create homepage and reference element containers in content classes such as Stylesheets, Header and Footer

Running Searches

Mon, 08 Jun 2009 17:46:36 -0500

The left hand menu of the CMS interface has a "Search" feature which can be quite powerful for running reports.

Recently we created a new TS website for our unit. The old ts webpages were moved to http://old.ts.vcu.edu.

The new TS website was considered a "soft launch" and was running side by side with the old site as content was moved. I needed to run through the entire project and identify where editors had linked to the old.ts.vcu.edu site for evaluations on what content was missing and most prominent, as well as make sure these links were removed when we canned the old.ts.vcu.edu site completely. This is an example of where the search feature is particularly useful.

I choose "search" in the left menu, then in my criteria I put in the old.ts.vcu.edu address. 67 page results were returned with the last editor and original editors eID so that I simply have to contact them and tell them which pages to update.

Multiple Publishing Targets

Mon, 18 May 2009 16:01:44 -0500

At our university we use different servers for database driven content and server side scripting. For instance pubinfo.vcu.edu is often used to reference faculty directories. But in any website, you may find that you need to publish pages in your project to two or more different URLs.

To do this you will create different "Pubishing Targets". Though you have two different Publishing Targets, you will only be able to create one "Publishing Structure". You can combine the structure for both in the one tree. Add all the possible folders to the structure, remembering that "root" is not specific to any URL.

Publishing target example:
URL 1: http://www.foo.com/blah/stuff/
URL 2: http://www.cnn.com/stories/

Publishing Structure example:

root

blah

stuff

stories
images
documents

In the Project Structure, define your publication packages using one of the two Publishing Targets. Be sure to also set the Publishing Structure inside the HTML variant folders.

*Be careful to not overwrite your publication packages. If you are using multiple publication targets, use "inherit publication package" carefully.

Rad Text Editor

Mon, 04 May 2009 17:39:22 -0500

With the upgrade from 7.5 to 9.0 a new text editor was introduced.

Known issues include:

a space before the text in table cells
numeric bullets in text editor appear as 1-1 instead of 1-a, but render correctly on page
text editor renders differently in IE8, includes a text height button which may be confused for headers, to use a header style look under the "Normal" drop down
undo does NOT undo last action, but all the changes since the last save (auto save every 30 seconds)

The new text editor adds a nice interface for uploading images and connecting to links.

Please note: if you are having image alignment issues, it's most likely a result of styles in your CSS, this also goes for table borders and padding which are determined by the global stylesheet not user imputed values

Exceptions for Particular Page

Mon, 27 Apr 2009 17:25:27 -0500

As part of Best Practices, use as few foundation/Master Pages as possible for a project.

For one particular site, I was delighted to see all the pages looked like they could be accommodated by one single template. Except one page. This index page did not display the headline of the page, where all the other subpages did. The absence of this one element meant I needed to change my entire template structure, or build a new template just for the index page.

Ha - Noway!

Instead I made a template exception using render tags and the page GUID.

<%hdl_title%>

I created an info placeholder "GUID of Page" to pull the page GUID to valuea. Then I compare it to the GUID of the index page. I found the GUID of the index page by looking under "Page Information" in SmartEdit. GUIDs (Gloabl Unique Identifier) are assigned to pages, elements, and content classes. RQL uses GUIDs to handle passing commands through the API. In this case, I am assigning the index page's GUID to valueb and comparing valuea to valueb.

If valuea equals valueb then write nothing:

Else write headline