Release Notes for Ultraseek 5.6
-------------------------------
February 2006
This file contains the release notes for both Ultraseek 5.5 and Ultraseek 5.6.
Additional product information is available at http://www.ultraseek.com/
Ultraseek is now a product of Autonomy Corp. http://www.autonomy.com/
New Features
------------
Ultraseek now supports PDF 1.6 (Acrobat 7) documents.
Ultraseek can now index documents from Sharepoint websites.
Ultraseek can now index Microsoft Outlook mail messages.
Implementation Changes
----------------------
The default installation directory on Windows is now:
C:\Program Files\Ultraseek\
The default data directory on Windows is now:
C:\UltraseekData\
Bugfixes
--------
[BZ0680] Collections modified via XPA did not automatically start
[BZ0838] Merge collection meta-data would still refer to previously deleted collections.
[BZ1309] NTLM Domain was referenced in the Help file
[BZ1319] Admin UI: Top Requested Documents report shows URL (%) encoded terms
[BZ1322] Response cache associated misspelled term with correct term results
[BZ1323] Start Page Expert matches at
tag, not beginning of HTML document
[BZ1341] Web Services error in "search" (Error: tuple[1] not integer)
[BZ1344] search from XPA v2.0 and earlier generates seek.error: tuple[1] not integer
[BZ1358] the HTTP server wasn't percent-decoding the "key" part of a URL Query string
[BZ1363] show_topic_counts incorrect when style setting included "all terms must match".
[BZ1365] In passage-based summaries, use case-insensitive matching for lower-case query terms
Release Notes Archive
---------------------
To see cumulative changes since an earlier release of Ultraseek,
visit the release notes archive on the Ultraseek web site:
.
Patchnotes for 5.6
------------------
Initial release: February 2006
Patchnotes for 5.6.1
--------------------
Released: April 2006
Bugs Fixed:
----------
[BZ0256] AdminUI: "View Sites" could display a URL multiple times.
[BZ0867] Handle SSCE Error during spell suggestions
[BZ1016] URLs containing adjacent slashes (//) are now sent to the remote
document server.
[BZ1285] File scanning collection did not follow symlinks to directories.
[BZ1310] Retain the word scores checkbox setting on the advanced search page.
[BZ1324] "query_each" failure is now reported as "query too long"
Several edge-case bugs with too-long queries are also fixed.
[BZ1325] Titles extracted from HTML or generated from URLs were present
in BIF, caused problems with K2 indexing.
[BZ1336] K2 batch full exception was not handled in some cases.
[BZ1375] Highlighting could split HTML or XML character entities, causing
pages to be not well-formed.
[BZ1380] Inconsistent decisions could be made by the duplicate-URL preference
algorithms - resulting in a different duplicate-URL being selected
as the preferred URL every time the spider visited a duplicate.
[BZ1388] Nulls in plain text caused the parser to give up early.
[BZ1403] Remove newlines from hrefs when parsing HTML documents
Patchnotes for 5.6.2
--------------------
Released: August 2006
Bugs Fixed:
----------
[BZ1369] Minor syntax error in ultraseek.wsdl.
[BZ1372] Remove copyright page expert feature is too aggressive.
[BZ1423] Fields parsed from XML files contain an additional space.
[BZ1431] Security vulnerability with logfile.txt.
[BZ1432] The smart summarizer is too aggressive when truncating long
sections of punctuation.
[BZ1438] Ultraspider bug - Delete URL throws an error when the URL
contains a space.
[BZ1444] NTLM Authentication failed if the Domain required Unicode.
[BZ1445] Spidering through a proxy server using NTLM does not work.