Sun Apr 13 18:13:28 1997  Andrew Scherpbier  <turtle@kiwi>

	* htlib/strcasecmp.cc (mystrncasecmp): Fixed it.  Only one string
	was converted to lower case.

	* htlib/URL.cc (parse): Added extra check so that a colon is only
	recognized if it comes right after the machine name.  It used to
	just look for a colon anywhere in the path.  This broke some bogus
	URLs.

Wed Jan 15 21:42:36 1997  Andrew Scherpbier  <turtle@kiwi>

	* htdig/HTML.cc (do_tag): Applied patch by William Miller that
	will fix a problem with the <base> tag.

Sun Jan 12 20:34:50 1997  Andrew Scherpbier  <turtle@kiwi>

	* htlib/URL.cc (URL): Implemented David Filiatrault's suggestion
	when dealing with anchors in URLs.  Anchors are now removed
	without affecting anything after the anchor (like CGI parameters).

	* config.html: Finally completed this file.  It now describes the
	basic files needed to run the search engine.

	* attrs.html: Fixed some problems pointed out by Darren Maglidt.
	Missing attribute documentation for keywords_meta_tag_names was
	added.

	* htsearch/Display.cc (setVariables): Applied patch by Sander van
	Zoest which fixed a problem with the use of the method_names
	attribute.  The docs say that individual parts can be quoted while
	the code didn't allow it.

Sat Dec 14 09:21:25 1996  Andrew Scherpbier  <turtle@kiwi>

	* htdig/Plaintext.cc (parse): Applied patch by "Miller, William G"
	<William.Miller@unisys.com> to fix a problem in the PlainText
	parser of htdig.  Before, if the plaintext parser was called, it
	used its contents literally.  Now it replaces '<', '>', and '&'
	with the corresponsing SGML entities.  Now, when htsearch displays
	the excerpt for a match, it will actually show these symbols.

	* htlib/Dictionary.cc (Get_Next): Applied patch by "Miller,
	William G" <William.Miller@unisys.com> to fix the Dictionary
	class.  The Start_Get() and Get_Next() dictionary traversal
	routines were severely broken.  Thanks William!

Sun Dec  8 21:27:09 1996  Andrew Scherpbier  <turtle@kiwi>

	* htsearch.html: Updated docs to reflect the new keywords
	value.  Also added a the MODIFIED output template variable to the
	list of variables.  (I forgot to do this before...)

	* htsearch/htsearch.cc (main): Added support for the new
	"keywords" HTML search form value.  This attribute can be used to
	add hidden terms to a search.  The terms will not show up in any
	of the display methods, nor will they be hilighted.  The keywords
	are added, separated with "and".

	* htlib/URL.cc (URL): Added suggested patch by Brian H. Powell.
	This patch allows URL of the form http:./foo.html
	I am not sure if this really is legal, but it won't hurt anything
	else.  (What if the document the above URL is in came from a
	different service like ftp?  Is it now supposed to switch services
	while keeping the same path?)

	* htsearch/Display.cc (display): Added suggested patch by Esa
	Ahola.  This will make htsearch only load document information for
	matches that are actually going to be displayed.  Before, it would
	get document information for all matches.  This will have to be
	reworked as soon as we need to add some ranking rules that depend
	on the document information.

Sat Nov 30 13:15:59 1996  Andrew Scherpbier  <turtle@kiwi>

	* htsearch/parser.cc (perform_or): Added a patch by GLen Pringle
	(pringle@cs.monash.edu.au) that fixed a bug in the boolean
	expression parsing.  If an expression ended with an operator (AND
	or OR), the parser would dump core.  The patch added error
	checking and recovery.

	* htdig/SGMLEntities.cc (translate): Fixed a small problem which
	caused a compile error on some platforms.  Thanks Micheal Bass for
	finding and fixing this.

Fri Nov 29 12:23:36 1996  Andrew Scherpbier  <turtle@kiwi>

	* htdig/HTML.cc (do_tag): Fixed a bug in the <a> tag parsing which
	only looked at the first attribute.  If the tag contained both a
	name= and an href=, only the first one was seen.  Now they are
	both seen.
	(parse): Scott Laird found a problem whereby htdig would coredump
	under certain conditions.  He fixed the problem by properly
	parenthesizing the condition that was supposed to check for this
	problem.

	* htdig/Server.cc (robotstxt): Added support for the new
	robotstxt_name attribute and also fixed a bug which caused the
	parsing of the robots.txt file to produce wrong results.

	* htcommon/defaults.cc (defaults): Added the robotstxt_name
	attribute which is used to match the 'user-agent' lines in
	robots.txt files on servers.  This will allow different databases
	to be built depending on the robots.txt file and the name htdig
	advertises itself as.

Sat Oct 26 09:20:59 1996  Andrew Scherpbier  <turtle@kiwi>

	* htdig/HTML.cc (HTML): Added a new attribute,
	keywords_meta_tag_names, which contains a list of meta names for
	which the content will be used as document keywords.

	* htlib/Dictionary.cc (rehash): Fixed problem with the
	reallocation of the hash table.  The variable newCapacity was used
	to create the new array before it was actually computed, hence
	giving a random number and frequently causing virtual memory to be
	exhausted.

Tue Oct 22 08:45:50 1996  Andrew Scherpbier  <turtle@kiwi>

	* htlib/good_strtok.cc (good_strtok): Fixed problem with
	characters outside of the 0-127 character range.  Changed the
	indexing into the valids array to use unsigned char instead of the
	default signed char.

Sun Oct 13 9:57:47 1996  Andrew Scherpbier  <turtle@kiwi>

	* htcommon/WordList.cc (WordList): Changed the parameters to the
 	Dictionary constructor from (32767, 0.85) to (203, 10.0).  This
 	should severly reduce the demaind for memory in htsearch and
 	htdig.

	* htlib/Dictionary.cc (rehash): Changed the new capacity
 	calculation to allow for load factors greater than 1.0.  Also
 	changed the default load factor from 0.75 to 10.0 to reduce the
 	number of memory\ reallocations.

Tue Sep  3 16:25:01 1996  Andrew Scherpbier  <turtle@kiwi>

	* htsearch/parser.cc (parse): Added support to ignore words when
	parsing the boolean search expression.

	* htsearch/htsearch.cc (setupWords): Added check for the
	bad_words. They are now marked to be ignored so that the
	expressions parser can do just that...

	* htdig/HTML.cc (HTML): Added a new configuration attribute
	minimum_word_length which has a default of 3.  This determines the
	minimum length of words that are indexed.  (Kinda follows from the
	name, doesn't it?)

Mon Sep  2 13:17:34 1996  Andrew Scherpbier  <turtle@kiwi>

	* htcommon/WordList.h (class WordList): Changed the words
	Dictionary to be a pointer instead of an automatic so that the
	dictionary can be initialized with a much larger table size and
	also a higher threshold.

	* htcommon/defaults.cc (defaults): Added .ram to the list of bad
	extensions.  .ram files are realaudio files which I don't think
	I'll be able to write a parser for since it is a proprietary
	format!  :-)

	* Makefile.in (install): Fixed a bug that caused the rundig script
	to be generated incorrectly.  The symptom was that the endings
	databases were recreated every time rundig was run.  They only
	need to be created once.

	* installdir/rundig: Added support to run htfuzzy on the synonyms
	database.  This goes hand in hand with the changes to htdig.conf.

	* installdir/htdig.conf: Added extensive comments to the example
	configuration file.  Also added some more example attributes,
	including exclude_urls, max_head_length, and search_algorithm.  I
	hope this will make the configuration of htdig a little easier.

	* htdig/Document.cc (getdate): Darren Maglidt
	<dmaglidt@horizons.com> helped find a problem with the getdate()
	method.  It turns out that on SunOS localtime() and timegm()
	cannot be used together.  I changed it to use localtime() and
	timelocal() instead.

	* htmerge/words.cc (mergeWords): Put the string "htmerge:" in
	front of all verbose output produced by the htmerge program.

	* htdig/HTML.cc (do_tag): Fixed a problem with the patterns being
	matched.  It was using the Compare() method of the StringMatch
	class but since both "a" and "area" were supposed to be matched,
	the <AREA> tag was never seen because it was matched with the "a"
	pattern.  By changing this to use the new CompareWord() member of
	StringMatch, this problem went away.

	* htlib/StringMatch.h (class StringMatch): Added members
	CompareWord(char*) and CompareWord(char *, int&, int&) to
	complement the Compare(char*) and Compare(char*,int&,int&)
	members.

	* htlib/strptime.cc (mystrptime): Changed strptime.c to
	strptime.cc and renamed the function from strptime to mystrptime.
	This is to finally get rid of all the problems with machines that
	have a strptime but no prototype.  Now mystrptime will ALWAYS be
	used, ignoring the system version.

	* htlib/lib.h (mystrptime): Added prototype for the new mystrptime
	function.

Mon Aug 26 22:30:55 1996  Andrew Scherpbier  <turtle@kiwi>

	* htsearch/Display.cc (generateStars): Added suggested code by
 	Nathan Neulinger <nneul@umr.edu> to fix possible problems with
 	divide by zero.

Thu Aug 22 12:33:54 1996  Andrew Scherpbier  <turtle@kiwi>

	* htdig/Retriever.cc (Initial): Fixed problem with multiple start
 	URLs.  It now actually *really* works!  The problem was that I
 	used strtok() in a loop while the URL class also uses it to parse
 	URLs.  Gotta love side effects!

	* htmerge/words.cc (mergeWords): Applied patch supplied by Andy
 	Hooper <hooper@knot.queensu.ca> to prevent the removal of files if
 	/bin/sort fails for some reason.  Also added support for the
 	TMPDIR environment variable to set the temporary directory
 	/bin/sort should use.

	* htsearch/htsearch.cc (setupWords): Added explicit support for
 	ISO-Latin-1 characters within words that are searched for.

	* htdig/HTML.cc (do_tag): Added support for the parsing of the
 	<frame> tag withint HTML documents.  The 'src=' attribute is the
 	same as the 'href=' attribute in an <a> tag.  (This was suggested
 	by Darren Maglidt <dmaglidt@horizons.com>)

Thu Aug 15 19:50:55 1996  Andrew Scherpbier  <turtle@kiwi>

	* htdig/Document.cc (Retrieve): Fixed a problem I introduced when
	making the document retrieval more efficient.  I used
	Connection::read_partial() instead of io::read()...  Since the
	headers of a document are read one line at a time using the
	buffered I/O or the io class, there was still data in the internal
	buffer that the Connection::read_partial() didn't know about.
	(ARGH!  Took me a while to find this one!)

Wed Aug 14 09:26:52 1996  Andrew Scherpbier  <turtle@kiwi>

	* htdig/Retriever.cc (parse_url): Applied small patch supplied by
 	"Brian H. Powell" <brian@natinst.com>

Tue Aug  6 12:45:44 1996  Andrew Scherpbier  <turtle@ender>

	* htdig/SGMLEntities.cc: Fixed the SGML entities table.
	Appearantly, the original document I go the info from had all the
	grave and acute accents switched around.  Thanks "Patrick
	A. Morin" <pmorin@clic.net> for pointing this out.

	* htdig/Document.cc (Retrieve): Changed the document retrieval to
	be more efficient.  It now no longer reads documents one line at
	the time.  The only possible disadvantage is that now '\r'
	characters will be part of the contents.  (This is a good thing
	for non-HTML files!)
	Thanks Chung-chieh Shan <ccshan@husc.harvard.edu> for pointing
	this out.

	* htdig/HTML.cc (do_tag): Fixed bug that caused the 'keywords'
	META tag name not to be recognized.  Thanks Anat Rozenzon
	<Anat.Rozenzon@telrad.co.il> for discovering the problem.
	(parse): Fixed bug with the SGML entities.  A ';' was added to the
	end of translated entities.

Sun Aug  4 13:56:06 1996  Andrew Scherpbier  <turtle@ender>

	* htnotify/htnotify.cc (send_notification): Added check for null
	subjects.  This caused htnotify to segfault.
	(send_notification): Added the use of the htnotify_sender
	configuration attribute instead of the hard coded htdig@sdsu.edu
	(oops!)

Tue Jul 30 08:58:06 1996  Andrew Scherpbier  <turtle@ender>

	* htsearch/htsearch.cc (main): Added a small patch by Nathan
 	Neulinger <nneul@rollanet.org> to check for the existance of the
 	CONFIG_DIR environment variable which will be used instead of the
 	compiled CONFIG_DIR constant.  This allows a small shell script to
 	set the CONFIG_DIR environment variable to what ever and allow one
 	copy of htsearch to use multiple configuration directories.

	* htsearch/Display.cc (createURL): Modified the createURL method
	so that the value of the SCRIP_NAME environment variable is not
	pushed through the URL encoder.  It was reported that the '-' in
	cgi-bin was incorrectly converted to %2d which caused problems for
	some server/browser combinations.

Tue Jul  9 14:04:11 1996  Andrew Scherpbier  <turtle@max.sdsu.edu>

	* Makefile.in (install): Modified the install procedure to reflect
	the new files and images that are required now.

	* htlib/String.h: modified the '<<' operator to support unsigned
	char as an argument type.  This caused problems after adding the
	international language support.  If an unsigned character was
	used, it was silently upgraded to an int.  (You gotta hate c++!)

Wed Jul  3 08:45:39 1996  Andrew Scherpbier  <turtle@max.sdsu.edu>

	* htsearch/Display.cc (setupImages): This method will setup things
	so that the star image URL can depend on the URL of the match it
	is associated with.  This makes it easy to distinguish between
	matches in multiple databases, for example.

	* htsearch/htsearch.cc (doFuzzy): Rewrote the fuzzy algorithm
	handling so that it actually does what the documentation said.
	The correct weight is now assigned to each word produced by a
	fuzzy algorithm.  In order to make everything more consistent, the
	Exact fuzzy algorithm was added.  It is there only to make the
	code cleaner.

	* htfuzzy/Synonym.h (class Synonym): Added a new fuzzy algorithm:
	Synonym.  It allows for the mapping of words to other words.  No
	default synonym table is given because I couldn't find a good
	one.  :-(

Tue Jul  2 13:11:30 1996  Andrew Scherpbier  <turtle@max.sdsu.edu>

	* htdig/HTML.cc (parse): Added support for SGML entities through
	the use of the new SGMLEntities class.  This means that things
	like '&amp;' will be translated to '&'.  Look in SGMLEntities.cc
	for a table of the known entities that will be translated.

	* htfuzzy/EndingsDB.cc (expandWord): Fixed a small bug which
	caused rules that needed to remove characters to remove an extra
	character. This meant that something like 'supply' was expanded to
	'suppied' instead of 'supplied'.

	* htfuzzy/Endings.cc (getWords): Made sure that if the word is not
	a root, the root is added to the list of words.  (It wasn't
	before...)

Mon Jun 24 14:00:01 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htmerge/words.cc (mergeWords): Got rid of the mmap() call to
	read in the whole wordlist file.  This was eating up LOTS of
	virtual memory and caused problems on some systems.  Thanks to the
	suggestion of Nathan Neulinger <nneul@umr.edu> I replaced the
	whole mmap() and qsort() sequence with a simple call to the unix
	sort program.  There appears to be no difference in speed between
	the two ways of doing the sorting.

	* htmerge/htmerge.cc: Split the source up into docs.cc and words.cc

Wed Jun 19 15:40:48 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htsearch/htsearch.cc (main): Added support for the
	'matches_per_page' attribute.  This is set with the '-p' option.

	* htcommon/defaults.cc (defaults): Added the 'matches_per_page'
	attribute which determines what its name implies...  It is a
	default only and can be overriden in htsearch.

	* htsearch/htsearch.cc (main): Added '-x' option to do the
 	opposite of the '-l' option.  It sets a pattern for URLs to
 	exclude from the search.

Tue Jun 18 09:16:51 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig/Server.cc (push): Fixed bug with the robots.txt matching
	algorithm.  It now actually works...  :-(

Mon Jun 17 17:14:31 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig/Document.cc (Document): Removed a redundant constructor
	for the Document class.  Now one constructor can deal with a URL
	if supplied.
	(Document): Added an optional max_size parameter to the
	constructor to set the max_doc_size.  This is needed in case the
	config value for "max_doc_size" is less than a robots.txt file.
	This caused the robots.txt file to be truncated for certain
	servers.

	* htdig/Retriever.cc (got_href): For verbose level 1, changed the
	display of '.' to either '-' or '+' depending on whether the URL
	was accepted or rejected.

Wed Jun 12 13:27:28 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig/HTML.cc (do_tag): Added support for the 'keywords' meta
	attribute in addition to the old 'htdig-keywords' attribute.  This
	is because some other search engines also seem to use this method.
	(do_tag): Added support for the 'htdig-index' meta attribute.
	This can be used to start indexing again after indexing was turned
	off with the 'htdig-noindex' attribute.

Tue Jun 11 10:03:48 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htsearch/display.cc (excerpt): Replaced the constant text for
	when there is no excerpt with the attribute value
	'no_excerpt_text'.

	* htfuzzy/Metaphone.cc (addWord): Fixed bug in the metaphone
	algorithm which could cause a segmentation fault in certain
	cases.

	* htdig/Retriever.cc (Retriever): Changed the constant weight
	factor for normal text from 1 to whatever is specified in the
	'text_factor' attribute.

	* htcommon/defaults.cc (defaults): Added a 'text_factor' attribute
	and set the default to '1'.  This will allow the administrator to
	set the weight for text that isn't in any special tag.

Thu May 23 17:57:51 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig/Document.cc (Retrieve): Added the 'Referer:' header to
	HTTP requests if that information is available.

	* htcommon/defaults.cc (defaults): Added 'max_doc_size' attribute.

Thu Apr 18 09:44:59 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htlib/Configuration.cc (Read): Multiple lines of the
	configuration file can now be combined by placing a backslash as
	the last character of a line that is continued.
	This works the same as line continuation in Makefiles.

	* htdig/Retriever.cc (Initial): Added support for multiple start
	urls.

Wed Apr 17 17:22:11 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htlib/ParsedString.cc: Changed the parsing to allow for
	the inclusion of files by using back-quotes around a filename.
	The filename may also contain variables.
	The result of this is that configuration files can now ``include''
	text files in attribute values.

Thu Apr 11 10:08:00 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig/Document.cc (Retrieve): Added support for the HTTP proxy
	protocol.  If the 'http_proxy' attibute is defined, it will be
	used to retrieve documents.

	* htcommon/defaults.cc (defaults): Added default value for
	http_proxy attribute.

	* htwrapper/htwrapper.cc (main): Changed the location of the
	configureation file from DATABASE_DIR to CONFIG_DIR.  Thanks to
	Dale Bingham (dale@bitshop.com) for pointing this out.

Wed Mar 20 13:41:01 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htsearch/htsearch.cc: Changed all references to the gdbm library
	to the GDBM_db class instead.

	* htlib/GDBM_db.h (class GDBM_db): Added this class to encapsulate
	the GDBM library functions.  It is derived from the Database class.

	* htsearch/htsearch.cc (main): htsearch will no longer segfault
	when no search words are specified.

	* htdig/Parsable.cc (setContents): Fixed memory leak.  The
	contents String was not released before a new one was created.

Thu Mar  7 14:54:37 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htsearch/htsearch.cc: Added a new option '-r' which will cause
	htsearch to produce raw output which is easy to interpret with a
	wrapper program.  This is so that some other program can take the
	search results and format it the way it wants to (or do other
	things with it...)  The format of the raw output is defined in the
	documentation.

	* htfuzzy/htfuzzy.cc: Added support for the creation of the
	endings database.  This used to be done with a perl program but is
	now done with htfuzzy when specifying the 'endings' algorithm on
	the command line.

	* htcommon/defaults.cc (defaults): Added three new attributes:
	endings_affix_file, endings_dictionary, and excerpt_show_top.
	Check the documentation for the meaning.

Fri Feb 16 17:13:40 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htlib/Connection.cc (assign_server): Fixed problem with 64 bit
	machines like dec alpha.

Thu Feb 15 14:36:58 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig/HTML.cc (do_tag): Changed the way URLs are seen.  A '?'
	will now also terminate the URL.  This is to prevent htdig from
	going down infinite virtual web trees.

Mon Feb 12 16:46:00 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htcommon/defaults.cc (defaults): Changed the values of a lot of
	the variables to use the new variable expansion feature of the
	Configuration class.  The new variable database_dir now defines
	the directory to use for everything else.  database_base is now
	defined in terms of database_dir.
	The old way of defining database_base will still work as it used
	to in previous versions.

Sun Feb 11 16:04:04 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htsearch/htsearch.cc (setupWords): Added support for the
	'metaphone' and 'endings' fuzzy algorithms.

	* htfuzzy/Metaphone.cc: Add the 'metaphone' fuzzy algorithm.
	It is used in the same way as the 'soundex' algorithm.

	* htfuzzy/Endings.cc: Added the 'endings' algorithm which uses a
	dictionary of word roots and all legal endings for those roots.

	* htfuzzy/createDict.pl: Added this program to create
	two dictionaries to be used with the 'ending' fuzzy algorithm.
	The input to this program is a language definition plus a
	dictionary.  These files can be gotten from the Ispell program or
	other sources.  Only a small English dictionary is included, but
	any size and number of dictionaries can be used, including
	dictionaries in other languages.

Wed Jan 31 20:45:58 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htmerge/htmerge.cc (Merge): Added better statistics reporting
	output.  All output lines now start with "htmerge:".

Fri Jan 26 11:28:06 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* examples/README: Added some more example files to the directory
	and the README.

	* contrib/htwrapper/htwrapper.cc (main): Took out the HTML output
	for the header and trailer.  This means that the header and footer
	HTML now has to be placed in the files defined by the
	'search_results_header', 'search_results_footer', and
	'nothing_found_file' attributes.
	This was done to allow more flexibility in things like background
	color/image for the search results.

Tue Jan 23 14:09:50 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* configure.in: Added test for the need for a prototype for the
	gethostname() library call.

	* htlib/Connection.cc: Added test for the need for a prototype for
	the gethostname() library call.

Thu Jan  4 09:22:44 1996  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig/Server.cc (Server): Fixed bug in the robots.txt retrieval
	code.  Previously, if a whole server was protected, htdig would
	get a authorization failure on robots.txt and it would interpret
	that as meaning that the server was unreachable.  (Chuck Lewis
	found this one...)

Fri Dec 29 16:41:59 1995  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig: Added support for Basic authentication with HTTP
	requests.  The '-u' option specifies the username and password to
	use.

Wed Dec 20 09:28:27 1995  Andrew Scherpbier  <andrew@sdsu.edu>

	* htcommon/WordList.cc (BadWordFile): Fixed nasty bug with empty
	lines to the bad_words file.  (found by Craig.)

	* htsearch/htsearch.cc (htsearch): Can now do proper fuzzy
	searching.  The weight factors for each of the algorithms is
	calculated in to the overall weight computation for matches.

	* htdig/HTML.cc (parse): Added code to detect HTML quoting
	escapes.  They start with '&' and end with ';'.  If these
	characters were seen as valid punctuation, the name of the quoted
	entity would become part of the previous word if there was no
	space.  This happened with "Billy&quot;" which was seen as
	billyquot.

Sun Dec 17 17:29:28 1995  Andrew Scherpbier  <andrew@sdsu.edu>

	* htlib/Match.h (class): Added int hasPattern() member to
	determine if a pattern was compiled with a Match (or derived)
	object.

	* htsearch/htsearch.cc (main): Added the ability to specify more
	than one limit pattern.  They are now all put into a StringMatch
	object for efficient matching.  Multiple limit patterns can be
	specified either by repeating the '-l' option or my separating the
	patterns with '|'.

	* htfuzzy/*:  Added fuzzy index creation.  Updated configure.in
	and Makefile.in to reflect the addition of this subdirectory.

Mon Dec 11 10:58:12 1995  Andrew Scherpbier  <andrew@sdsu.edu>

	* htmerge/htmerge.cc: Added better error recovery in case or all
	of the needed files are unavailable.

Sun Dec 10 13:14:30 1995  Andrew Scherpbier  <andrew@sdsu.edu>

	* htsearch/htsearch.cc: Added better error recovery in case some
	or all of the needed files are unavailable.

Sat Dec  9 16:16:43 1995  Andrew Scherpbier  <andrew@sdsu.edu>

	* htdig/Retriever.h and htdig/HTML_Parser.h: made the destructors
	virtual since the classes contain other virtual methods.  (Caused
	compiler warning.)

	* htdig/Document.h and htdig/Document.cc: used "enum
	Document::DocStatus" which some compilers barf on.  Took out the 'enum'.

Tue Nov 14 14:33:28 1995  Andrew Scherpbier  <andrew@sdsu.edu>

	* Finished port to IRIX 5.3

	* htlib/htlib.a: Added strptime.c.  It will be used if the system
	doesn't have it.

