Appliance Update: texis-7 and texisScripts-8.1.2

Post Reply
User avatar
Kai
Site Admin
Posts: 1271
Joined: Tue Apr 25, 2000 1:27 pm

Appliance Update: texis-7 and texisScripts-8.1.2

Post by Kai »

Two new appliance updates are available: texis-7 and texisScripts-8.1.2. Both must be updated together (i.e. they are co-dependent). Customers with the Catalog module should not apply this upgrade, as it also requires a catalog upgrade (coming out soon). Fixes in texis include:

Results Authorization against a web server with URLs with Content-Encoding or Transfer-Encoding could fail with out of memory errors
Add more Maximum Process Sizes between Large and Huge. Add a note to not use Unlimited.
Fixed issue where web server (vhttpd) sub-processes may hang indefinitely, causing access problems to the appliance (web requests timeout/hang). Usually caused by signal, e.g. client aborts the connection.
Caveat: Cookies with Path values that are not a prefix of the URL that is setting them will now be accepted during crawls, as most browsers do. May allow an application to alter the cookies of another application that is on the same server, which may be a security issue. Configurable with the Strict Cookie Paths setting
Cookies without a leading period in the Domain attribute are now accepted from servers, as per RFC 6265
Fixed issue where some crawled pages' JavaScript strings caused the error "Will not allocate 2147483882 bytes of memory: JavaScript exceeded scriptmem limit" (where the byte count is of the form 0x8000.... or 0xc000....)
List/Edit URLs with a leading-wildcard query (e.g. *.pdf) were returning all URLs. Affects Texis versions since 2012-04-24.

Fixes and features in texisScripts-8.1.2 include:

Added metasearch Max Backend Data Size setting
Fixed Modify Parametric Fields' "apply at next 'new' crawl" and later force migrating field information loss
Fixed some foreign language documents generating invalid titles
Fixed fatal profile errors in metasearch backends preventing further profiles from returning results
Fixed Walk Status possibly failing to find data if loaded while walk is shutting down
Fixed crawls improperly allowing long file extensions
Binary files now able to properly use http headers in Data From Field sources
Added ability to include internal settings in backups (for tech support troubleshooting)
High verbosity crawls now log more reasons for link rejection (meta robots nofollow, max depth, etc)
Links excluded by Max Pages, Max Depth, and Max Bytes are no longer saved in todo
Links excluded due to Max Pages, Max Bytes, or user-cancelled walks are now logged properly with high verbosity
Fixed potential rare error condition when changing multiple settings while simultaneously changing replication settings
Streamlined loading of All Walk Settings and Search Settings pages
Increased efficiency of crawling large files
Byte counts now use pre-modification (keep/ignore tags, etc) sizes
Added breadcrumb links to Single Profile Disk Space page, and links to non-standard subdirs (if any)
Added <ResultDecl> for Additional Fields in XML search results
Renamed 'Tech Support Information' to 'System Information'"
Fixed rare condition where Walk Status could temporarily show walk not running immediately after launching
Disallow profile names containing colon or backslash
Admin SOAP getSettings()/setSettings() now provides better SOAP Failure errors
Fixed URLs with no category getting a random category.
Fixed potential duplicate BestBet results from multiple profiles on the same metasearch backend
Crawls now allow https content by default
Fixed crawler repeating URLs with ignorecase and skipping placeholders
Fixed 'Match Info' presenting invalid links with Query Logging
Added 'Category' as a Data From Field target
Post Reply