Several Appliance updates are now available: texis 6.00.129..., texisScripts-7.1.8, applianceman-7.1.8, and license-... The license update is a prerequisite for the texis update; it does not otherwise change license terms.
The texis update (a prerequisite of the texisScripts update) contains the following fixes:
* Texis Version 6: official release of many pre-existing features
* Improved abstracts
* Setting Max Page Size to -1 may have failed and resulted in `Bad parameter in the function TXfetchSetMaxDownloadSize' error message
* Some RTF documents were producing little or no text (but copious <META NAME>s instead) when run through anytotx plugin
* Some `w/N' withinmode fixes
* Invalid-schema <timport> run after a valid <timport> was causing ABEND (may have affected replication)
* ABEND in jQuery ui.core.js fixed (javascript crawls)
* Memory leak in fetch lib fixed; should reduce memory use during long-running crawls
* Crawling an FTP site may hang and consume CPU, usually after an extended period, especially during a directory request.
* Capture occasionally missed meta data from MSWord DOC files.
* Support \uNNNN escapes beyond \uFF, in JavaScript
* Try to avoid `No forward progress' error from some sites that return extra data after Content-Encoding gzip stream; now reports `Extra NNN bytes input ignored'
* Avoid `Cannot decode deflate data: zlib returned Z_DATA_ERROR: incorrect header check in the function TXzlibTranslate' error from some HTTP/1.1 fetches of Content-Encoding: deflate URLs
* Open Document Format documents supported
* Avoid ABEND for certain SQL math expressions such as `pow('x'*'y', 0.43)'
* `Cannot convert charset ... to ... via converter ...: Timeout' error was possible during crawls/<fetch>, after long pause, especially for custom small Vortex <fetch>es
* Erroneous HTTP/0.9 requests with POST data can cause web server to timeout (e.g. if Appliance was crawling itself)
* An HTTP/1.1 fetch of gzip+chunked page can cause a fetch timeout sometimes
* `oe'/`OE' ligature (U+0153/U+0152) was not being expanded under `loose' Character Match Mode
The texisScripts update contains the following:
* Changed XSL rendering engine. If this causes any problems, Profiles can change "XSL Engine" from Internal to Legacy until issues are resolved.
* Apply Apperance" and "Test Appearance" checkboxes are removed, updating "test" or "test & live" settings is now accomplished by the submit buttons "Update Test" and "Update Live and Test", respectively.
* Refreshing pages whose content didn't change will now properly follow links that weren't already crawled
* "Allow XML" has been removed
* Added category data to "Show Parents" xml output
* Fixed SOAP output for results containing Set parametric values
* Dataload SOAP WSDL now includes parametric fields set to not output
* "Number" type parametric fields are now "String" in the WSDL, as they may be empty, which many APIs attempt to parse as invalid doubles.
* Added ParametricQuery to SOAP interface
* Added ModifiedDateGreaterThan, ModifiedDateLessThan, and SiteQuery to soap input & output, increased xmlOutputVersion to 1.06
* Query Highlighting setting now split into Results Highlighting and Context Highlighting settings
* Restore display of result Url when right best bets are turned on.
* Fix order of best bets in meta search results.
* Fixed ordering of XML elements with no results and TopBestBet matches
* Catalog imports now send mail to addresses listed to "Notify" upon completion
* Added Exclude Extensions setting
* Fixed validation issues in the RSS feeds produced by the search interface
* docs dataloaded without a Modified time will now have their Modified set to 'now'
* CSV data files now supported for Catalog import
* Added more options to Data from Field sources: URL Path, URL Query Var etc.
* Fixed jump values for PreviousLink through SOAP interface
* Single-value set parametric fields can now be dataloaded using the non-set syntax
* Errors when sql inserting for catalog imports are now logged specifically for that record, rather than being printed to the walk status log with little context
* Try to avoid `Trying to insert duplicate value ... in index xerrorurl.btr' errors
* Fix for metasearching profiles with a "Set" type Parametric Field
* Replication failed to send to a receiver with ACLs enabled
* Added "API Logging" to allow logging of SOAP & dataload request XML & response XML for troubleshooting
* Added missing field data for Modified, Visited, & Charset to <ThunderstoneLinks> Result elements
* Fixed 'XSL Fix' page output
* Added profile-specific and system-wide replication logging
* List/Edit page display now detects and displays mismatches between expected Parametric Fields and database table's contents
* Fixed links to file shares with multibyte international characters
* Fixed mounting file shares with multi-byte foreign language file names on older appliances
These updates are available via the Maintenance -> Check for Updates menu.
The texis update (a prerequisite of the texisScripts update) contains the following fixes:
* Texis Version 6: official release of many pre-existing features
* Improved abstracts
* Setting Max Page Size to -1 may have failed and resulted in `Bad parameter in the function TXfetchSetMaxDownloadSize' error message
* Some RTF documents were producing little or no text (but copious <META NAME>s instead) when run through anytotx plugin
* Some `w/N' withinmode fixes
* Invalid-schema <timport> run after a valid <timport> was causing ABEND (may have affected replication)
* ABEND in jQuery ui.core.js fixed (javascript crawls)
* Memory leak in fetch lib fixed; should reduce memory use during long-running crawls
* Crawling an FTP site may hang and consume CPU, usually after an extended period, especially during a directory request.
* Capture occasionally missed meta data from MSWord DOC files.
* Support \uNNNN escapes beyond \uFF, in JavaScript
* Try to avoid `No forward progress' error from some sites that return extra data after Content-Encoding gzip stream; now reports `Extra NNN bytes input ignored'
* Avoid `Cannot decode deflate data: zlib returned Z_DATA_ERROR: incorrect header check in the function TXzlibTranslate' error from some HTTP/1.1 fetches of Content-Encoding: deflate URLs
* Open Document Format documents supported
* Avoid ABEND for certain SQL math expressions such as `pow('x'*'y', 0.43)'
* `Cannot convert charset ... to ... via converter ...: Timeout' error was possible during crawls/<fetch>, after long pause, especially for custom small Vortex <fetch>es
* Erroneous HTTP/0.9 requests with POST data can cause web server to timeout (e.g. if Appliance was crawling itself)
* An HTTP/1.1 fetch of gzip+chunked page can cause a fetch timeout sometimes
* `oe'/`OE' ligature (U+0153/U+0152) was not being expanded under `loose' Character Match Mode
The texisScripts update contains the following:
* Changed XSL rendering engine. If this causes any problems, Profiles can change "XSL Engine" from Internal to Legacy until issues are resolved.
* Apply Apperance" and "Test Appearance" checkboxes are removed, updating "test" or "test & live" settings is now accomplished by the submit buttons "Update Test" and "Update Live and Test", respectively.
* Refreshing pages whose content didn't change will now properly follow links that weren't already crawled
* "Allow XML" has been removed
* Added category data to "Show Parents" xml output
* Fixed SOAP output for results containing Set parametric values
* Dataload SOAP WSDL now includes parametric fields set to not output
* "Number" type parametric fields are now "String" in the WSDL, as they may be empty, which many APIs attempt to parse as invalid doubles.
* Added ParametricQuery to SOAP interface
* Added ModifiedDateGreaterThan, ModifiedDateLessThan, and SiteQuery to soap input & output, increased xmlOutputVersion to 1.06
* Query Highlighting setting now split into Results Highlighting and Context Highlighting settings
* Restore display of result Url when right best bets are turned on.
* Fix order of best bets in meta search results.
* Fixed ordering of XML elements with no results and TopBestBet matches
* Catalog imports now send mail to addresses listed to "Notify" upon completion
* Added Exclude Extensions setting
* Fixed validation issues in the RSS feeds produced by the search interface
* docs dataloaded without a Modified time will now have their Modified set to 'now'
* CSV data files now supported for Catalog import
* Added more options to Data from Field sources: URL Path, URL Query Var etc.
* Fixed jump values for PreviousLink through SOAP interface
* Single-value set parametric fields can now be dataloaded using the non-set syntax
* Errors when sql inserting for catalog imports are now logged specifically for that record, rather than being printed to the walk status log with little context
* Try to avoid `Trying to insert duplicate value ... in index xerrorurl.btr' errors
* Fix for metasearching profiles with a "Set" type Parametric Field
* Replication failed to send to a receiver with ACLs enabled
* Added "API Logging" to allow logging of SOAP & dataload request XML & response XML for troubleshooting
* Added missing field data for Modified, Visited, & Charset to <ThunderstoneLinks> Result elements
* Fixed 'XSL Fix' page output
* Added profile-specific and system-wide replication logging
* List/Edit page display now detects and displays mismatches between expected Parametric Fields and database table's contents
* Fixed links to file shares with multibyte international characters
* Fixed mounting file shares with multi-byte foreign language file names on older appliances
These updates are available via the Maintenance -> Check for Updates menu.