Release Notes: The default settings now enable CompactIndex, IncrementalCitations, UNICODE, and "fast clones lookup" features. A "MinFixedPatternLength" parameter was added to searchd.conf, and has a default value of 6. &Nbsp; entity processing was added (non-standard but used on some pages). Update instructions were added to the README. Bugs regarding stopwords used in OR or NOT expressions were fixed. A configuration bug in fast clone lookup was fixed. s.cgi error messages are no longer reset if include is used in s.htm. Sending of ER_TOOSHORT was fixed in the case where ER_NOQUOTED in searchd. Incorrect processing of HTML entities in the non-Unicode version was fixed.
Release Notes: A bug in reverse citations merging which was introduced in 1.2.9 was fixed. Incorrect redirects merging was fixed. Incorrect merging of the reverse citation index was fixed in the case when URL A refers to B which redirects to C and A also refers to C and one of references from A disappeared. The fact that the ResultsPerPage parameter from s.htm overrides PS cookie was fixed. Small improvements and fixes were made to init.d/aspseek, aspseek-mysql-postinstall, the spec file, and the manual pages. The DBLibDir directive was added to aspseek.conf and searchd.conf.
Release Notes: This version added support for the HTTP POST method in s.cgi, IncrementHopsOnRedirect and RedirectLoopLimit options and sed-like \1 to \9 sequences in the Replace command in aspseek.conf. Tag parsing was improved in the index to handle omitted quotes, and the option to limit results by a range of dates was fixed. Several rare core dumps in searchd and s.cgi and two rare memory leaks in searchd were fixed. FreeBSD portability fixes were made, along with fixes and improvements in the results cache and man pages.
Release Notes: A generic client library (libaspseek) and module for Apache server, new man pages (index(1) and aspseek-sql(5)), fixes for 64-bit platforms such as Alpha for gcc3/ISO C++ conformance, sped-up citation merging, fixes for incorrect (in some cases) processing of -u -t -s options in "index", a fix for a bug with not closing file in "searchd", a fix for a bug in excerpts processing, a mechanism to reuse URL IDs (saves memory in "searchd"), and a fix for a bug that caused "orphaned" URLs in database.
Release Notes: The package was made more portable, and it should now compile on non-Linux systems. The code is compilable with gcc3. Linking on certain Linux platforms was fixed. Checking of deleted entries in the inverted index was fixed. Processing of clones was fixed. The random number generation facility in s.cgi was fixed. CheckOnly and CheckOnlyNoMatch behavior was fixed (GET was used instead of HEAD). More langmaps for Czech charsets were added, and another (smaller) list of Italian stopwords was written.
Release Notes: This release implements a buddy-like heap and buffered file in "index" for better memory usage and faster processing. It fixes improper clones processing in "index", a few coredumps in "searchd" and "index", a few memory leaks in "index", and the -P and -A flags in "index." The amount of stack used by "index" has been reduced. Concurrency between searchd threads having many DB connections has been improved; Options have been added to "index" to delete URLs from an inverted index and to re-create broken citation files from the DB. The I/O nce of "searchd" has been optimized.