Smartsite now supports the outerelm viper parameter for assisting the compliance layer by telling it what the element name is returned by the viper.

When security information is written to the RAW channel output, the information is based on id's of the usergroup, accesstype and accesslevel. However, when synchronizing this information (or using export/import) to another site/database, the usergroup-id's and/or the accesstype- & accesslevel-id's may not be equal.

Therefore, two extra attributes have been added to the security information: usergroupcode and accesslevelname.

The Importer will use these two attributes (if present), to map the usergroup-code to the corresponding usergroup-id, and to map the accesslevelname to the correct accesstype and accesslevel.

When synchronizing items, including security information, between two sites (either using Outscaling or export/import), just make sure the usergroup-codes and the name column in accesstypelevels correspond on both ends. In that case, it doesn't matter whether or not they have the same primary key.

To eliminate the ambiguity between self-closing and empty <se:member> elements, the value attribute can be used, just like in the se:text macro.

When more than one version of a translation was specified, the argument signature could not be different for each version, or arguments could not be retrieved by name. This caused unexpected behavior.

The Scf macro replaces Behavior and is much more powerful. See documentation for Scf.

Use the new parameter FormatSourceId instead.

SqlQuery can now execute queries without validating the parameters by setting the parametervalidation parameter to relaxed.

ConnectionBuilder crashed when the clipboard was locked. A message is now shown that the clipboard is locked.

Performance counters were lazy initialized causing performance counter not be available at all times. Now site is started, performance counters are available.

When a placeholder had missing dependency placeholders, the render error that occurred could not be trapped anywhere but in the root SXML element.

The placeholder.directives macro is used at the top of render templates to add directives such as those for ASP.NET (<@ page ... >)

The placeholder.headsection macro renders custom html head data, not rendered by any of the mandatory placeholders.

The placeholder <se:placeholder.directives /> can be used to add extra directives or comments to the rendered output and should be placed above the HTML open tag.

The placeholder <se:placeholder.headsection /> can be used to add special (meta-) tags to the head section and should be placed immediately after the title element.

The Smartsite iXperion 1.1 update script adds these directives to the ELBA elements baselayout() and init().

Intellisense in the SXML Editor in the TestSuite and the CMS showed local Viper methods that were marked as 'Deprecated'.

Friendly url creation was suppressed when using the extension Viper methods.

You can now get and set the response encoding directly using the Response.Encoding() viper method.

Attributes weren't recognized correctly. Font size was set without unit (should be %). Multiple css classes on a single element were separated by commas in stead of spaces.

The BBDecode method no longer supports the font tag.

A BBEncode method is introduced that will re-encode xhtml to bb code, if possible.

When selecting a xsl file within the Import Xml File(s) wizard to perform a xsl transformation prior to importing the selected xml file(s), the wizard did abort with the error "type mismatch".

Smartsite advises placing all users and visitors in the Guests group, to make administration of public areas of the site easier.

Smartsite did not this notation http://www.mysite.com/news/ but does now.

If a global translation and a local translation existed with the same name, a reference to the translation in the same folder as the local translation would resolve the global instead of the local.

CSS detection of raw string data in CMS items could fail, which could then lead to scanning errors due to nesting of CDATA sections.

The then, else and case elements in the If and Switch macros now support the 'rem' attribute for remarks.

When started as test server, the TestSuite could still produce an error message that it could not be started as a Publication Server when the configured test server address was in use.

Due to an invalid temp path used by Outscaling, synchronizing two cms instances fails when using Windows Server 2008.

When using Windows 2003, a call to check whether the illegal path exists or not just returns False. However, when the same call is executed running Windows Server 2008 this causes an "Illegal path" exception.

Since the call is executed within the initialization code of the standard Outscaling webservice (wsOutscaling), Smartsite cms is unable to create the wsOutscaling webservice object and therefore none of the methods provided by this webservice can be executed. Effectively, this means that Outscaling between two cms instances is not possible.

To fix this problem, the illegal path has been replaced with a valid path.

The authentication in iXperion using the Login macro did not use the path and domain settings in the web.config, causing authentication not to work across multiple host headers in the same site.

The StateType CTFP is introduced. It is used on the Nr field of the new SML (Smartlet) ContenType to set a .NET CLR Type used for state persistence necessary for Smartlets. See the Smartlet Architecture documentation.

The Outscaling dialog Synchronize Items now supports the option (when the logged in user does have admin rights) to select a (custom) view. When this option has been selected, all items returned from the selected view will be synchronized to the selected target server. Use with care!

This versions automatically refreshes all user views and regenerates the schemabound views vwFriendlyname and vwTsContent.

Execute this procedure after each schema change in the Smartsite CMS database!

The stored procedure pr_RebuildView is no longer needed and can be dropped.

test4

To distinguish between content that may be sent to (one of) the subscribers and content which must be sent (automaticallly from Outscaling's background component), the Outscaling ServerInstance configuration now includes two views:

1. Content View - "master" content view which is used to determine if an item may be sent to a subscriber; this view is used e.g. when using the manual action Synchronize Items as well as within the "missing parent" mechanism.

2. AutoSync View - view which is used (if properly configured) by Outscaling's background component. Only items included within this view will be automatically synchronized to a subscriber. Of course, items to be sent must be included within the specified content-view as well.

The Outscaling Synchronize Files dialog, using the "Synchronize a complete folder" option and the "Use pulling" (sub)option, failed with an "type mismatch error".

Scf being built upon jQuery, the body onload attribute should be removed and the call to any dynamically included Javascript statements should be configured in the Smartsite.config placeholder.javascript macro (resultformat): $j(onLoad);

These translations can be accessed using the vipers in the locale namespace, such as locale.getstring() and locale.translate().

The FriendlyName field would be empty (because the original would share the same name), but the permanent bit is also set in the FriendlyNameType, disabling the FriendlyName textbox.

In some cases, hyperlinks in the returned data could not be corrected.

When NAP pages are requested using the classic Url pattern, with a querystring Id parameter, Smartsite will internally reroute the page using the Virtual Path Provider and the friendly url (with ASP.NET extension).

Direct NAP preview is not possible, because ASP.NET has to be in control, accessing Smartsite iXperion using the Virtual Path Provider. However, a preview of NAP pages in the CMS did not cause an error to be shown to the user.

When a page was given a friendly Url with the 'Folder' flag set (no file extension), the NAP page would not be recognized by ASP.NET. The Url is now internally rewritten with a '.aspx' extension.

A request with no document in the Url could not be mapped to ASP.NET using the built-in VPP Url rewriting. For instance,  /?id=NAP_PAGE would not work, where  /smartsite.net?id=NAP_PAGE did.

When restoring the emptysix site (database, site files, registry), it is best practice to immediately rename the site.

This includes renaming the site root folder on disk, as well as the site's registry root and adjusting several addresses etc. within the configuration files (of course, using the Smartsite Config Editor).

However, there's also one important Outscaling configuration setting within the database which needs to be adjusted as well and is often overlooked.

This setting is the Url property of the server instance used to configure the Publication engine.

To make configuration a little easier, Outscaling now supports a specific value for the Url property of a specific server instance.

Just use the value "%PubInteropServiceAddress%" (without the quotes) as Url property for the pub server instance (as is done within the EmptySix 1.1 template site), and Outscaling will automatically substitute it with the correct endpoint adress of the InteropService service of the publication engine (as configured within the web.config).

A hidden input was generated without a closing element tag.

Browse access inheritance is now handled by a database trigger, in stead of CMS code. Access is inherited directly from the parent folder for items as well as folders.

The Browsable bit that is added by the update script is exposed in the properties dialog in Manage Contenttypes.

Edit Channel now validates that the given DefaultDocument is unique within the site.

The translation name now has to start with a letter, followed by 2 to 79 letters, digits and underscores, ending with a parameter block. The parameter block consists of a pair of parantheses with any number (including none) of comma separated parameter names (starting with a letter, followed by up to 79 letters, digits and underscores).

If no parameters are needed and the parentheses are omitted, an ace rule will add them for the user.

Previously, the publication server loaded all channels, even channels that weren't active. If these inactive channels were invalid, the publication server would throw an exception. Inactive channels are nog ignored. They are excluded from the Channels view.

Object reference not set  in ConfigEditor when there is no IISMetaPath.

Compliance configuration in the empty site was updated to conform to xhtml strict standard.

A new webcontrol has been added to the library, with the progid SixWebtoolkit.SmartletListInput.

This webcontrol can be used within the cms manager to allow an editor to select/configure one or more Smartlets.

To use this webcontrol, you should add a new contenttypefield (e.g. "SmartletZone") to the desired contenttype and bind it to a text field, e.g. ctspecifictext1. Then add the ctfp "EditProgId" with the value "SixWebtoolkit.SmartletListInput" to this contenttypefield.

The separator which should be used between each smartlet call can be set using the ctfp "ctrl:SmartletDelimiter". Defaults to an empty string.

When an Url Command (?hid=xxx;value=....) is passed over the querystring, the url.commands.get() method does not urldecode the values.

The '&' Querystring separator were hidden because the '&' sign represents an accelerator in Windows.

The Viper DataTable.Rows.RemoveRange is broken and does not work as expected. The startindex was 0-based instead of 1-based and the count parameter acted as the endindex instead of the actual number of rows.

When an exception occurred during the scan for unresolved vipers in the final stage of the rendering, these errors were thrown. The errors are now hidden away from the visitor.

Using the default formatting, this resulted in an extra outer ul element, making the output non-compliant with xhtml. If the output has been fixed by dropping the ul element from the resultformat, the new version will have no outer element, once more breaking compliance. To revert to the old behaviour set release compatibility to 1.02 or below.

Underscores are now supported in sql parameter names in the SqlQuery macro.

When opening the manager, you are now also automatically logged in on the publication server. Smartsite creates a single sign on session that you can use to to render items on the publication server on behalf of the user that is logged in in the manager. This single sign on session is limited to the browser session in which the manager is opened.

If all children of a folder are hidden, displaying siblings on one of these items would result in a Null Reference Exception.

When using levelformat, the last row was not sent to the output when it had level 0.

Several fixes and adjustments have been implemented within contenttype(s) export/import, as well as copy/paste contenttype functionality, since they both use the same underlying technique:

- Code field of usergroups is now written within the contenttype xml (export/copy) and used during import/paste to match usergroups by code.

- After processing (import/paste) the usergroups information within the contenttype definition, a check is performed whether or not the usergroup with the code ADMINISTRATORS has access to the created/updated contenttype. If not, this contenttype-usergroup relation is automatically added.

- Mandatory contenttype-channel relation for both RAW and AIM channel are checked. If they do not exists after processing (import/paste) the contenttype xml, they will be created automatically.

- Mandatory channelmappings (contenttype - channel - rendertemplate relation) for both RAW and AIM channel are checked. If they do not exists after processing (import/paste) the contenttype xml, they will be created automatically.

- After importing the "CtfLocked" contenttype field property, all subsequent contenttype field propererties of the same contenttype field fail with the message "Object variable or With block variable not set". This has been fixed as well.

The ContentTypeDefinition schema (which can be found within the ../system/schema folder of each site) has been changed, to reflect other changes ("browsable" field has been added to the contenttypes table and contenttype import/export now matches usergroups based on code).

The TestSuite could under circumstances enter a deadlock situation.

When there's a normal translation and one or more local translations with the same name, a local version of the translation could not be removed.

This was due to the fact that for each translation version (both normal and local), the AIM reference created was marked as "Hardlink".

This has been changed, now every local translation (when there are both normal and one or more local translations having the same name) will be marked as "Softlink" and only the normal translation will be marked as "Hardlink".

Regardless of the workflow being used, updating an existing item using the Import XML Files wizard will fail with the error message:

"Workflow error: no path from previous status ASSIGNED to current status ACTIVE."

This is/was due to the fact that each workflow didn't have an ExtendedAttribute child element (with name="TransitionMode") on the Activity element with name="Publish". This means the workflow engine can not directly make a state transition from ASSIGNED to ACTIVE.

By adding an ExtendedAttribute child-element with name="TransitionMode" and either Value="Optional" or Value="Connectable" to the Activity with name="Publish", this could be resolved. However, this will also result in a change of the statuses shown within the status dropwdown of Edit Item (Active will now be shown as well).

Since this is an unwanted side-effect, a new TransitionMode has been introduced: Hidden. When the transitionmode of an activity has been set to Hidden, the workflow engine will still be able to make a transition from this activity to the following activity. But, the status dropdown will not be affected, since the underlying code which determines the available statuses will not take a Hidden transitionmode into account.

The output decoding took place after the cache file was written, leading to incorrect cache files.

Returns the user id of the anonymous (Guest) user.

Viper String.ReadFromFile now removes the Byte Order Mark from the file when reading it.

Smartsite now correctly checks that the request was made on one of the ip addresses that are bound to this computer.

Errors page for exception caused by invalid urls and inactive items now have a much friendlier error page.

The metadata macro accidently activated the image server.

You can change the default character used as a word separator by setting the variable @whitespacechar in stored procedure pr_FriendlyNames.

After changing this value, execute pr_FriendlyNames 'PREP' to synchronize the Manager configuration with the new whitespace character.

Updates to the procedure pr_FriendlyNames will preserve th configured value for @whitespacechar.

The Importer has a built-in mechanism to choose an alternative contenttype (based on contenttype-inheritance) when an item within the import xml file has an unknown contenttype.

However, the Importer didn't give a warning when that situation occurred and just "silently" switched to an alternative contenttype.

Now, a warning is displayed within the progress dialog. Also, when importing a single file (and without using the bulk import option), the wizard will now also give a warning when the selected import xml file contains one or more unknown contenttypes.

The NoAim macro now always skips the rendering of its contents if an AIM render is performed.

When an error occurred within the AIM rendering, the error was not correctly returned.

The render template for the AIM channel must be changed, in order to return the correct error message.

The render template should look like:

<se:aim xmlns:se="http://smartsite.nl/namespaces/sxml/1.0"

error="{string.enclose(html.attributes(version,'1.0',encoding,response.encoding()),'&lt;?xml','?&gt;')}{html.tag(error, this.error.message(default='Unknown Error'))}"

>

<se:references aim="off" channel="DEF"/>

<se:aimscan/>

</se:aim>

Since the deafult rendertemplate for the AIM channel has been changed to return the underlying error details within the xml, the references tab is now able to display this error message when AIM rendering fails.

The AIM scan of an System Translation item failed to resolve the correct references present within the Replacement field, if there were multiple translations (both local and global) having the same name.

This was due to the fact that the wrong translation version was retrieved, which was then used to determine the references.

Because the Content-Length was already set, browsers were waiting for more data and therefore did not display the error message that was generated.

The repository database scripts are now compatible with the redesigned workflows.

Due to a missing check if it's an AIM request, rendering an ASP.NET Page on the AIM channel failed with the exception "ASP.NET rendering is only supported using the Smartsite Virtual Path Provider.".

To check whether a request is actually an AIM request, you can now use the viper method request.isaimrequest().

The Autonomous AIM flag (AIMSource.Autonomous registration type within vipers and AIMScanMode.Autonomous scanning mode within macro's) is only intended to be used within translations and rendertemplates.

Therefore, now an exception will be thrown when the Autonomous flag is being used outside translations and/or rendertemplates.

Of course, no exception should have been thrown, but the registration of the relation should have been skipped.

The manager now fully support IE8.

Set Workflow gave an internal error when attempting to set the workflow for deleted items.

To configure locales and translations.

Directory is now created automatically, if needed.

All query parameters are now encoded. A compatiblity setting is available to revert to previous behaviour.

Before invoking this method, make sure the source text does noty include any tags or vipers.

For best results, normalize the source string with string.normalizewhitespace().

This viper can be used to remove unresolved vipers from a string. Note that unresolved vipers can only exists in a string if you read or query item data in raw mode, or if a viper cannot be resolved due to incorrect parameters.

It now looks for the overload that requires the least number of explicit type conversions. In case different overloads require an equal number of conversions, the first is selected.

When AIM rendering of a specific item fails, the AIM scanner will now retry a failed item (default) three times. After (default) three consecutive failures, the AIMState of that particular item will be set to Error (=6) and the item will be skipped from further AIM scanning.

The AIMState Error will be reset to Dirty when a full rescan is requested, or when an item having the Error AIMState is being saved/activated.

The target of the transfer loads its own item data, but changes number and parent field to those of the source of the transfer. Calling the itemdata.field viper with the transfer source as reference would incorrectly return the itemdata of the transfer target.

Lookuplists already worked with drowdownlists (single value selectors) but now also work with duallists (multi value selectors). The sortindex is not yet respected. This will be fixed in the next release.

When a contenttype field uses a contentrelation to store related items, and one of the related items has a showfrom in the future, that item is not included in the RAW channel xml output of that field.

This was due to the fact that the active channel view was being used to format the content relations to xml notation, whereas the vwContent view should have been used.

The character set/content encoding of the response was not taken into account when decoding the returned result.

When the render service (which is used when previewing an item within the cms manager) encounters an error, e.g. because the sxml is invalid or inconsistent, the error is not displayed anymore within the preview tab.

Instead, a generic message "Render request to InteropService did not return a response or an error was raised." is displayed.

Webservice calls to Outscaling's external webservice would fail with the exception "Object reference not set to an instance of an object.".

This was due to a bug within the Smartsite.Interop.IIS.WebserviceHandler, which threw this exception when the Path registry setting within the site's logging configuration was missing.

As work-around, add the Path setting to the site's registry configuration, specifically the Logging configuration: HKLM\Software\Smartsite Software\Sites\[sitename]\Logging.

When you try to copy an item (which is a folder) to itself within the cms manager, an error dialog is shown with the message "source and destination can not be the same".

However, why shouldn't you be allowed to create a copy of a folder directly below itself? Well, there's no reason for it, so this check has been removed.

Due to a bug, the description of an item within a feed was set to null when that item didn't have a link field.

Due to a bug, the string resources for the internal default locale "en-us" were not loaded when the channel's default locale has been set to e.g. dutch.

Since the internal default locale is being used to lookup translations, locale.translate was unable to translate strings to e.g. dutch.

Set the registry value of Manager/EnableSSO to false to disable Single Sign On.

On some servers/configurations any call to Outscaling's external webservice fails. When using Fiddler to see the actual response, it turns out there's a problem creating SixTools.General:

"Unable to cast COM object of type 'SixTools.GeneralClass' to interface type 'SixTools._General'. This operation failed because the QueryInterface call on the COM component for the interface with IID '{8F2E594C-CA12-4396-A2B1-69B751E742FF}' failed due to the following error: Error loading type library/DLL. (Exception from HRESULT: 0x80029C4A (TYPE_E_CANTLOADLIBRARY))."

On some systems, this behaviour can be reproduced within a test aspx page which tries to create SixTools.General and then call the CreateNewGUID() method.

Switching to Asp Compat mode internally within the http handler solves the problem.

Outscaling mechanism to remove non-existing constraints or automatically send the record causing the constraint viloation when synchronizing records (e.g. user or usergroup) didn't work properly.

Also, the dataeditor meta xml files (the default set which comes with our "empty six site") have been adjusted to use this mechanism for the standard Smartsite "objects" (Users, Visitors, UserGroups, UserRoles, etc.).

The Hasp session for the cms manager wasn't terminated when an AppDomain recycle occurs. Although the session automatically terminates after a timeout of 12 hours or when executing iisreset, the site shutdown code has been adjusted to explicitly close the hasp session.

Single Sign On (SSO) is not supported under IE6. SSO can be turned off in the registry by settings the registry value of Manager/EnableSSO to false.

This error was shown due to a bug. This bug has been fixed.

When editing a channel configuration item within the cms manager and the publication engine receives the ContentChange.Activation event for this change, an AppDomain recycle is initiated. However, sometimes the hasp session for feature 3 (Site instance) isn't terminated when the AppDomain recycles.

This may lead to the fact that the Site is unable to (re-)start, because the limit on the number of allowed logins to feature 3 has been exceeded.

To resolve this issue, now the hasp session is terminated explicitly before initiating the AppDomain recycle.

When saving a new item using the Office Integrator (a.k.a. Remote Editing), the item isn't created with the specified title. Instead, the item is created with a title like "_tempXXX_import No Title 1", where XXX equals the item's number. This only happens when using Word/Office 2007.

This was due to the fact that Office 2007 doesn't include a title element within the head section when a document is saved as html.

Custom code have been added to fix this problem.

Users that have more than one role may not be able to add forum topics, even though one or more of their roles allow this. The role that is chosen for adding the topic is indeterminate.

This has been fixed in 1.1 build 3.

Note that the CmsUpdate macro is not officially supported and may change or be dropped in a future release.

Due to an older, which turned out to be an incorrect, implementation of Active Directory support within the cms manager, an user was unable to login into the manager when he/she first logged in on the front-end and vice-versa.

This was due to the fact that the implementation within the cms determined the wrong SID (Security Identifier) and therefore wasn't able to match the determined SID with any user/visitor.

This implementation has been fixed, and now both the cms and publication engine use the same SID values.

When an user is already registered with the wrong SID value, the SID value will automatically be updated to the correct value.

Notice that you have to make sure your web.config has been properly configured to use Windows Integrated Security:

<location path="cms">

<system.web>

<globalization requestEncoding="iso-8859-1"/>

<authentication mode="Windows" />

<authorization>

<allow users="*"/>

<!-- allow access for all users, the manager uses it's own implementation for authenticating users -->

</authorization>

<httpModules>

<clear/>

<!-- clear the inherited httpModules -->

<add name="WindowsAuthentication" type="System.Web.Security.WindowsAuthenticationModule"/>

<add name="Session" type="System.Web.SessionState.SessionStateModule"/>

</httpModules>

<httpHandlers>

<clear/>

<!--clear the inherited httpHandlers-->

<add verb="*" path="/cms/webservice.dws" type="Smartsite.Interop.IIS.WebserviceHandler,Smartsite.Interop"/>

<add verb="*" path="/cms/*.dws" type="Smartsite.Interop.IIS.Handler,Smartsite.Interop"/>

</httpHandlers>

</system.web>

</location>

The Active Directory Membership Provider for the publication engine and the Windows logonhandler for the cms engine, both supporting the use of windows integrated security, have been updated:

- They now both have a symmetrical implementation of the StoreFullyQualifiedLoginName option (this option wasn't supported by the logonhandler before).

- The StoreFullyQualifiedLoginName option now also takes effect when no default domain (the first entry within membershipMappings, if any) is specified.

- Switching between Smartsite standard security and Windows integrated security is now much easier, both the provider and the logon handler do now use multiple lookups to find the appropriate user (based on the SID, loginname or loginname@domainname).

The recommended Active Directory (AD) configuration includes creating global groups (e.g. "ad_webmasters", "ad_editors" and "ad_chiefeditors"), in which the users, dependent on their role within the organization, are distributed.

Besides these global groups, (at least) one domain local group must be created (e.g. "ixperionaccessgroup") in which the global groups must be added as a member (and no user will be a direct member of the domain local group). This domain local group must then be specified within the web.config as SmartsiteAccessGroup (Smartsite.ActiveDirectoryConfiguration section).

However, Smartsite's Active Directory MemberShip Provider didn't take parent groups into account. With the recommended AD config, you would get an access denied message when you try to login, because only direct group membership was resolved. So, although the user is member of e.g. "ad_webmasters", the membership provider didn't reckon the user is a member of the specified SmartsiteAccessGroup.

Smartsite's Active Directory MemberShip Provider implementation have been adjusted to take parent groups into account. In the above example, when an user is member of "ad_webmasters", he/she is also considered to be a member of "ixperionaccessgroup", since the "ad_webmasters" group is a member of this group.

When using an Outscaling scenario, in which one of the cms server instances has been configured to use pulling, the items being pulled fail.

This had to do with an encoding issue, the rendered result returned from a webservice method wasn't correctly encoded and appeared to include a bunch of question marks only. This has been fixed.

The WCF service methods all use a custom login method to be able to impersonate the current user. This mechanism fails when using Windows Integrated Security.

Specifically when the default provider has been set to the Active Directory Membership Provider, this mechanism has been adjusted.

The CMS Setup installs a new version of the Hasp License Manager: version 12.47.

This version solves the stability issues reported with earlier versions. Therefore, make sure the version of the License Manager is 12.47 or higher.

To inspect the Hasp version number, browse to the License Manager (http://localhost:1947) and select About.

Contenttype fields for which the ctfp ctrl:isBitPattern has been set, such as the FriendlyNameType field used in many contenttypes, fails to import correctly. The underlying value, as specified within the xml, is ignored. This occurs when manually importing an item containing this field, as well as in an Outscaling scenario.

This was due to the fact that the storagehandler(s) used for this type of field didn't have logic to detect and handle a value which represents a bit pattern. The SimpleLookupList and SimpleLookupStorage storage handlers have been adjusted to support bit patterns.

An exception was thrown in the CmsWebCacheDependency when the ExternalReferences table contains a relative url.

The addin will now allow selection of .doc, .docx and .odf files by default. This can be overridden using the FileFilter registry value in the key where the addin is configured.

When using the FileInput control within the cms, images (or other filetypes) added using this control will have an "absolute" path (it starts with "/", e.g. "/images/myimage.jpg").

Outscaling uses the external webservice to synchronize files which are included within the aim references.

However, the external webservice fails to resolve paths which starts with "/". This has been fixed.

Due to a flaw in session logic, using sessions could lead to random errors. This has been fixed.

AIM requests on an item caused the item to have sessions set to disabled when in the normal request {session.enable()} was called. When the AIM request was performed before a regular request was performed, an exception was thrown.

Passwords with a length that is a multiple of 4 do not work in the login macro.

When calling the response.redirect viper from a host header B (a non-default host header) resulted in a redirect to the default host header of the same channel.

To be able to synchronize Localized String Resources between multiple CMS instances, first an appropriate Outscaling Datasource must be created.

However, notice that vwLocalizedStrings must be used as datasource name and not the table LocalizedStringRecources itself.

Due to some issues, however, records added or changed to/within LocalizedStringResources failed to synchronize. This has been fixed.

Furthermore, the dataeditor meta xml file vwLocalizedStrings.xml should be changed, especially when you have configured an Outscaling scenario with two (or more) cms server instances: add an allowedit node with value set to 2 to the field with name="ResourceKey". (It shouldn't be allowed to change the ResourceKey for existing records.)

When you change the content access for specific items using the Manage Content Access action, these changes are not synchronized to (other) cms server instances.

The Manage Content Access dialog now sets the activation date to the current date/time for the (active) items on which the content access security is set, so Outscaling is able to synchronize these items to the other subscribed cms server instances, if any.

(As part of the synchronization process, the content access security information is serialized within the xml output of the RAW channel and deserialized and interpreted during the import process on the target server.)

Because of this, when moving an item (using cut/paste within the cms manager) from one folder to another folder, this change is not reflected on the cms subscriber(s) of the local server instance.

In previous versions Smartsite refused to start when unknown data types were used in the Smartsite database.

When changing the sort-order on a folder item to custom and re-arranging the order of its child items, this change is not reflected on the publication engine. This was due to the fact that for the child items no ContentChange.Activation event was generated, which is needed to synchronize the change notification to the publication engine.

When using an Outscaling scenario of two (or more) cms server instances, each with their own database, SubTree Management changes are not synchronized to the subscriber(s) of the local server instance.

The first reason for that is the fact that the Outscaling Datasource for the table SubTreeHierarchy was missing from the default set of Outscaling Datasources. If your site doesn't have a datasource for the table SubTreeHierarchy, you should add it.

Second, because of the database design of the SubTreeHierarchy table, the Smartsite built-in MetaBinder object needed to be adjusted, to be able to determine the appropriate key when a SubTree entry was added.

Tables in other schemas than the default schema are now supported. The Smartsite database tables must reside in the default schema. As of iXperion 1.2, using multiple database schemas will be more restricted than in 1.1.

Links with spaces weren't recognized as links and therefore not parsed.

When you have a large and complex Active Direcotry, the Active Directory integration with Smartsite iXperion can lead to a degraded performance.

To improve the performance, once logged in Active Directory users are now internally cached.

With the recently added feature of caching the AD/LDAP user objects after a successful logon, a bug was introduced which resulted in user information (such as group membership and primarygroup) not being updated anymore once the user has been added to the database. This has been fixed.

The publication engine already had code to disable the so-called File Changes Monitor, which is responsible for causing an appdomain unload when a folder is deleted.

The http handler for the cms manager now also includes code to disable this File Changes Monitor.

The Active Directory membership provider had a built-in default for searchUserQuery: "cn=Users".

When the membership provider must query for an user, this searchUserQuery is used to create a new DirectoryEntry, which is then used to execute the search request. So, by default, the membership provider would only find users which are member of "cn=Users".

And although the default searchUserQuery can be overridden (on the membership provider element) with an alternative "OU", the provider should be able to search for an user on the "root" object (and any level within the domain hierarchy).

The Active Directory membership provider has been changed and doesn't specify a built-in default for searchUserQuery anymore.

Therefore, if you don't specify a searchUserQuery within the membership provider element, search operations for an user are executed on the "root" object.

When you do specify a searchUserQuery, then this searchUserQuery is used to create an DirectoryEntry object and that object is used to execute the search request. This means, if you specify a searchUserQuery, all the domain users which must have access to Smartsite must be a (nested) member of the specified "OU".

When a site visitor (with visitor-role) is already present within the Users table within the Smartsite database, the caching of the Active Directory group membership for this user/visitor wasn't properly cached. As a result, for each request Smartsite did query the Active Directory to determine group membership for the user, which seriously slows down the overall performance.

This has been fixed, and several other fixes/improvements (which were already available for 1.2) have been implemented as well, such as trace logging, no internal default for searchUserQuery (see work item 6873), etc.

The unnecesssary cache could lead to unnecessary memory consumption.