CHANGES for MHonArc

    http://www.oac.uci.edu/indiv/ehood/mhonarc.html

Revision history for MHonArc.  Changes listed are brief.  Consult
the documentation for further information/clarification.  It is
possible that not all changes are documented here.

MM/DD/YY
=======================================================================
#######################################################################
=======================================================================
06/09/97	(2.1.0)

    Enhancements/Changes
    --------------------

	o  Added the following resources:

		DEFRCNAME	Base name of default resource file
		FOLUPBEGIN	Begin markup for message follow-ups
		FOLUPEND	End markup for message follow-ups
		FOLUPLITXT	Markup for follow-up list entry
		GZIPEXE		Pathname of gzip executable
		GZIPFILES	Gzip files
		GZIPLINKS	Add ".gz" to filenames in URLs
		HTMLEXT		HTML extension for files
		MSGBODYEND	Markup at end of converted message body
		MSGPREFIX	Prefix for message filenames
		REFSBEGIN	Begin markup for message references
		REFSEND 	End markup for message references
		REFSLITXT	Markup for reference list entry
		TSORT		Sort threads by date or message number
		TSUBSORT	Sort threads by subject

	o  Changes to mhexternal.pl filter:
	
	    -  Two new options: useicons and iconurl="url".
	    -  Added some more content-types to CTExt hash.

	o  Added filter for text/enriched data (mhtxtenrich.pl).

	o  Changed default setting for MIMEFILTERS.  The following
	   was added:

	    model/*:m2h_external'filter:mhexternal.pl
	    text/enriched:m2h_text_enriched'filter:mhtxtenrich.pl
	    text/richtext:m2h_text_enriched'filter:mhtxtenrich.pl

	o  Some code moved to new library files (mhopt.pl,
	   mhfile.pl).


    Bug Fixes
    ---------
    Problem:    An undefined subroutine error occured during an
		RMM operation when there are messages with non-ASCII
		encodings in message headers.
    Solution:   Proper libraries are now loaded for RMM operations.

    Problem:   	Subjects using the "... -Reply" convention started
		a new thread.
    Solution:  	Fixed typo in regexp.

    Problem:   	Nested anchor markup occurs on index pages where
		a message's subject contains a URL and the subject
		is used to link to the message page.
    Solution:  	Replaced default routine that converts subject text


=======================================================================
05/23/97	(2.0.1)

    Bug Fixes
    ---------

    Problem:   	Bogus empty entry shows up in the database.  Bogus
		entry appears if a duplicate message is detected.
		Potential visible problems of bogus entry are: mailto
		links in headers corrupted; a blank listing in the
		index.  Maybe other potential problems.
    Solution:  	There is a be a bug in some versions of Perl where a
		hash key gets added invalidly.  The section of code that
		this occurs has been modified to avoid the problem.
		For existing databases with a bogus empty entry,
		MHonArc now removes any empty key entries.
		However, it is recommed to rebuild the archive.

    Problem:   	Index pages are not properly updated if zero messages
		are in the archive (like through the -rmm option).
		Even -editidx cannot cause a proper update.
    Solution:  	Page count was set to zero when no messages exist.  This
		caused problems in conditional checks in the write index
		routines.  To fix, page count is forced to 1 if their
		are no messages.

=======================================================================
05/16/97	(2.0.0)

    Enhancements:

	o  Added TCONTBEGIN, TCONTEND, TINDENTBEGIN, TINDENTEND
	   resources.  They are provided to support the formatting
	   of threads that split across multiple pages of a
	   multipage index.

	o  Added "nonfixed", "maxwidth", and "keepspace" options
	   to text/plain (mhtxtplain.pl) filter.

	o  A directory file is now used for performing archive
	   locking.  This should provide better portability across
	   operating systems.

	o  Dervied files returned from filters can be directories.
	   If so, the any directory is completely removed if the
	   message is removed.

	o  The mhexternal.pl filter can now output files into
	   a subdirectory with the "subdir" option.  This may be
	   useful for those wanting to use the "usename" option
	   and avoid potential filename conflicts.

	o  The $NEXT...$ and $PREV...$ resource variables now
	   behave as one may expect.  I.e.  The are always relative
	   to the base sort method: date, author, subject, message
	   number.  The REVERSE resource has no effect on their
	   values.  Therefore, the variables no longer refer to
	   the actual physical listing order.  This new behavior
	   no longer requires a users to redefine button and link
	   resources for the sake of specifying REVERSE.

	o  "Resent-To" and "Resent-Cc" will now have e-mail
	   address hyperlinked.

	o  Unknown types processed by mhexternal.pl now default
	   to have a ".bin" extension.  This should cause HTTP
	   servers to treat the file as application/octet-stream.

	o  Standard input can be specified as the source of a
	   mailbox by using a lone "-" on the command-line.  Note:
	   You may need to specify "--" first to shut-off option
	   processing so the "-" is not interpreted as the start
	   of an option.

    Bug Fixes:

	o  Problem:   	Index pages with last page links are not updated
		      	when a new page is added.
	   Solution:  	All index pages are regenerated if a new page
		      	is added.

	o  Problem:   	"-idxsize 0" does not give unlimited page size.
	   Solution:  	Problem fixed and now works as documented.

	o  Problem:   	"-maxsize 0" does not reset archive with
			unlimited messages allowed.
	   Solution:  	Problem fixed and now works as documented.

	o  Problem:   	Non-ASCII text encodings (=?ISO-..?...) are not
		      	properly decoded when EDITIDX is set.
	   Solution:  	The charset filter libraries are now loaded
		      	for EDITIDX.

	o  Problem:   	MODTIME resource setting not remembered across
		      	archive updates.
	   Solution:  	MODTIME resource is now properly stored in the
		      	database.

	o  Problem:   	Specifying -rmm and -single caused confusion on
		      	what MHonArc does.
	   Solution:  	The case is properly handled, with -rmm taking
		      	precedence.

	o  Problem:	4 digit year not checked before 2 digit year
			when parsing dates.
	   Solution:	Fixed.

	o  Problem:	Some index pages in multipage indexes not
			updated when messages expire.
	   Solution:	Fixed; all pages are regenerated when a message
			expires.

	o  Problem:	Setting MONTHS and MONTHSABR had no effect.
	   Solution:	Fixed.

	o  Problem:	$TNEXTFROM$ actually gave the value of $NEXTFROM$.
	   Solution:	Fixed.

	o  Problem:	Numerical timezone offsets not handled properly
			when the offset contained non-zero minutes.  This
			caused message to be sorted wrong by date.
	   Solution:	Fixed.

	o  Problem:	Some characters were not properly escaped when
			specifying the "in URL" modifier to a resource
			variable.
	   Solution:	Fixed.

	o  Problem:	Thread (formatting) information lost when a
			thread is split across multiple pages.
	   Solution:	Fixed.  New resources (TCONTBEGIN, TCONTEND,
			TINDENTBEGIN, TINDENTEND) exist to preserve
			formatting of threads across pages.

	o  Problem:	If specifying the REVERSE, the next/prev
			links in messages do not point to the proper
			message as one may expect.
	   Solution:	This should be fixed by the new behavior of
			$NEXT...$ and $PREV...$ resource variables.


=======================================================================
02/28/97	(2.0.0 beta 3)

	o  Removed bogus call to output_db() when GENIDX is active.
	   The routine is not loaded when GENIDX is active and the
	   call causes Perl to abort.

	o  Duplicates of the OTHERINDEXES resource should no longer
	   occur.  Before, if the OTHERINDEXES resource gets set
	   during invocation, mhonarc would store duplicate values,
	   causing unnecessary work to be performed.

	o  Fixed problem with blank indexes being created for
	   OTHERINDEXES.  Apparently, there is a bug in Perl, and
	   a work-around fix has been done.

	o  Fixed bug where an extra TSUBLISTEND could get generated.


=======================================================================
02/17/97	(2.0.0 beta 2)

	o  Added <nosubsort> element and -nosubsort option.

	o  Changed M2H_DATEFMT envariable to M2H_GMTDATEFMT.

	o  Documented LOCKTRIES and LOCKDELAY resources.

	o  New resources:

		AUTHORBEGIN	- Markup for start of author group
		AUTHOREND	- Markup for end of author group
		AUTHSORT	- Sort messages by author
		DAYBEGIN	- Markup for start of day group
		DAYEND  	- Markup for end of day group
		IDXLABEL	- Label for main index
		MSGGMTDATEFMT	- Format of $MSGGMTDATE$
		MSGLOCALDATEFMT	- Format of $MSGLOCALDATE$
		SUBJECTBEGIN	- Markup for start of subject group
		SUBJECTEND	- Markup for end of subject group
		TIDXLABEL	- Label for thread index

	o  New resource variables
		
		$IDXLABEL$	- Label for main index
		$MSGGMTDATE$	- Date of message in GMT format
		$MSGLOCALDATE$	- Date of message in local format
		$TIDXLABEL$	- Label for thread index

	o  Some enhancements to the documentation.

	o  Some code changes.


=======================================================================
01/31/97	(2.0.0 beta 1)

	o  RFC 1522 support.  Charset filters can be plugged in to
	   handle character sets not recognized by mhonarc.
	   The library, iso8859.pl, contains charset filters for
	   character sets iso-8859-[1-10].

	o  Multi-page index generation.

	o  Subject-based threading is done inconjunction with explicit
	   threading.

	o  Missing messages in a thread are recognized when processing
	   messages that utilize the References field.

	o  New resources for customizing thread indexes.  There is
	   now complete user control on how thread indexes are
	   formatted.

	o  Resource are available to access the next and previous messages
	   in thread from messages.

	o  Inclusion of files containing resource settings can be
	   done within a resource file.

	o  Support for a default resource file.  The default resource
	   file is one of the following, in the order of precedence:

		1. File specified by the M2H_DEFRCFILE envariable
		2. $HOME/.mhonarc.rc
		3. <Install-Library-Dir>/.mhonarc.rc

	   "<Install-Library-Dir>" is the path of the MHonArc library
	   directory defined during installation.  For MS-DOS, replace
	   ".mhonarc.rc" with "MHONARC.RC".

	o  Support for user defined resource variables.

	o  Support for user defined derived files for messages.

	o  Support for the Content-Length field.

	o  Modification time of messages, and derived files, can be
	   set to the Received/Date time specified in the message.

	o  More control on how mail headers get formatted.

	o  Added filter in mhexternal.pl to be the default filter
	   for main content-types: application, image, video, audio.

	o  List of enhancements to mhexternal.pl:
	
		- Show content-type if the data is not recognized.
		- New `ext' filter option to define the extension to
		  use for the created file.
		- New `type' filter option to define the "english"
		  type name of the create file.
		- Content-Disposition is supported.  Filenames are
		  ignored still unless `usename' argument passed to
		  filter.
		- Added recognition of few more content-types

	o  Updated date parsing to figure out the date of a message to be
	   more robust on how it parses a date.  If unable to parse
	   date, the message will take on the current time for sorting
	   purposes.  Also, character case is ignored in month and weekday
	   names.

	o  The comment declaration "<!--X-Body-of-Message-->" is
	   now inserted between the message header and body.  Helps
	   in the building of some search indexes to restrict searches
	   on message data.

	o  Many resource settings are no longer stored in database if
	   the resource is the default value.  Save some disk space and
	   allows resources to self adjust when a dependent resource is
	   changed.

	o  The text/plain filter in mhtxtplain.pl has the following
	   enhancements:
	   
	       -  Check for charset to control character conversion.
	       -  Integrated iso-2022-jp filter (keys off charset)
	       -  Filter option "quote" cause quoted text in message
		  to be italicized.
	       -  Filter option "asis" defines a list of charsets to
		  not convert to sgml entitites.  Example usage:

			asis=iso-8859-1:iso-8859-2

	o  mhtxt2022.pl has been removed since the code has been
	   integrated into mhtxtplain.pl.

	o  Resource file elements that have textual content (ie. no
	   line oriented content) can specify the "chop" attribute
	   to have the last end-of-line stripped from the content.
	   Example usage:

		<nextbuttonia chop>
		[Next]
		</nextbuttonia>

	o  Fixed bug in creating links of message-ids.  Mhonarc blindingly
	   made links of message-ids when editting messages w/o
	   consideration that the message-ids may already be linked.  This
	   caused markup like the following to occur:

	    <A HREF="..."><A HREF="..."><A HREF="...">...</A></A></A>...

	   Browsers handle the invalid markup with no problems, causing
	   the bug to go unnoticed for a long time.  Now, only new
	   added message-ids are scanned for when creating links.

	o  $readmail'FieldSep should now be used instead of $FieldSep
	   for separating duplicate fields in a parsed message header.

	o  The -scan output now prints a 4 digit year.

	o  Bogus space no longer appears in subjects and dates.

	o  Outdir permissions are not checked if -single specified.

	o  Some internal changes to how data is stored in databases
	   (needed for 1522 support).  v2.0 will automatically
	   modify 1.x databases to the 2.0 format.

	o  The -single option utilizes the same mail output routine
	   utilized by regular archive processing.

	o  MhonArc will now handle numbers with leading zeros when
	   the -rmm option is specified.

	o  New resources:

		CHARSETCONVERTERS
				- Specify character set filters
		CONLEN		- Honor content-lengths
		DECODEHEADS	- Decode 1522 encoded data, set for decode
				  only, as message headers are read (see
				  note below)
		DEFINEVAR	- Define resource variables
		DEFINEDERIVED	- Define user defined derived file
		EXPIREDATE	- Message cut-off date
		EXPIREAGE	- Time in seconds from current if msg expires
		FIELDSBEG	- Begin markup of converted mail header
		FIELDSEND	- End markup of converted mail header
		FLDBEG		- Begin markup of mail header field text
		FLDEND		- End markup of mail header field text
		FOLREFS		- Print links to explicit follow-ups & refs
		GMTDATEFMT	- Format of $GMTDATE$
		HEADBODYSEP	- Markup between converted mail header & body
		IDXPREFIX	- Prefix for multi-page main index filenames
		INCLUDE		- Read resources from other files
		LABELBEG	- Begin markup of mail header label
		LABELEND	- End markup of mail header label
		LOCALDATEFMT	- Format of $LOCALDATE$
		MAIN   		- Create main index
		MHPATTERN	- Expression for mesg files in a directory
		MODTIME		- Set file times to message dates
		MONTHS		- Full month names: EOL or ':' separated
		MONTHSABR	- Abbrieviated month names: EOL or ':'
				  separated
		MULTIPG		- Create multi-page indexes
		NEXTPGLINK	- Link to next page in main index
		NEXTPGLINK	- Inactive link to next page in main index
		NOCONLEN	- Ignore content-lengths
		NODECODEHEADS	- Leave message headers "as is" when read
				  (see note below)
		NOFOLREFS	- Do not print links to follow-ups & refs
		NOMAIN 		- Do not create main index
		NOMODTIME	- Do not set file times to message dates
		NOMULTIPG	- Do not create multi-page indexes
		PREVPGLINK	- Link to previous page in main index
		PREVPGLINKIA	- Inactive link to previous page in main index
		SUBJECTHEADER	- Markup for subject header in converted mail
		TIDXPREFIX	- Prefix for multi-page thread index filenames
		TSUBLISTBEG	- List begin in sub-thread
		TSUBLISTEND	- List end in sub-thread
		TSUBJECTBEG	- Begin markup for subject-based sub-thread
		TSUBJECTEND	- End markup for subject-based sub-thread
		TSINGLETXT	- Markup for mesg not part of a thread
		TTOPBEGIN	- Begin for top of a thread
		TTOPEND		- End for a thread
		TLINONE		- Markup for missing message in a thread
		TLIEND		- Thread idx list item end
		TNEXTBUTTON	- Thread next button template
		TNEXTBUTTONIA	- Inactive thread next button template
		TNEXTLINK	- Thread next link template
		TNEXTLINKIA	- Inactive thread next link template
		TNEXTPGLINK	- Link to next page in thread index
		TNEXTPGLINKIA	- Inactive link to next page in thread index
		TPREVBUTTON	- Thread previous button template
		TPREVBUTTONIA	- Inactive thread previous button template
		TPREVLINK	- Thread previous link template
		TPREVLINKIA	- Inactive thread previous link template
		TPREVPGLINK	- Link to prev page in thread index
		TPREVPGLINKIA	- Inactive link to prev page in thread index
		WEEKDAYS	- Full weekday names: EOL or ':' separated
		WEEKDAYSABR	- Abbrieviated weekday names: EOL or ':'
				  separated


	   NOTE
	       1522 processing is done when creating HTML output and
	       the (relevant) data stored in the database stays in
	       encoded form.  The DECODEHEADS resource can be set
	       to decode decode-only charsets when message headers
	       are read.  Hence, the decode-only charsets will be
	       stored in decoded form.	Regular 1522 processing is
	       still done to still-encoded data when generating output.
	       The default is NODECODEHEADS.

	o  List of removed resources:

		NOTSUBSORT, TSUBSORT

	o  New resource variables:
	   (NOTE: Some variables are only valid in certain contexts)

		$FIRSTPG$	- Filename of first page of main index
		$IDXPREFIX$	- Prefix to main index pages' filenames
		$LASTPG$	- Filename of last page of main index
		$NEXTPG$	- Filename of next main index page
		$NEXTPGLINK$	- Link to next page of main index
		$NUMOFPAGES$	- Total number of pages in index
		$PAGENUM$	- Current page number of index
		$PREVPG$	- Filename of previous main index page
		$PREVPGLINK$	- Link to previous page of main index
		$TFIRSTPG$	- Filename of first page of thread index
		$TIDXPREFIX$	- Prefix to thread index pages' filenames
		$TLASTPG$	- Filename of last page of thread index
		$TNEXTBUTTON$	- Button for next mesg in thread
		$TNEXTFROM$	- From of next mesg in thread
		$TNEXTFROMADDR$	- From address of next mesg in thread
		$TNEXTFROMNAME$	- From name of next mesg in thread
		$TNEXTLINK$	- Link for next mesg in thread
		$TNEXTMSG$	- Next mesg filename in thread
		$TNEXTMSGNUM$	- Next mesg number in thread
		$TNEXTPG$	- Filename of next thread index page
		$TNEXTPGLINK$	- Link to next page of thread index
		$TNEXTSUBJECT$	- Next mesg subject in thread
		$TPREVBUTTON$	- Button for prev mesg in thread
		$TPREVFROM$	- From of previous mesg in thread
		$TPREVFROMADDR$	- From address of previous mesg in thread
		$TPREVFROMNAME$	- From name of previous mesg in thread
		$TPREVLINK$	- Link for prev mesg in thread
		$TPREVMSG$	- Previous mesg filename in thread
		$TPREVMSGNUM$	- Previous mesg number in thread
		$TPREVPG$	- Filename of previous thread index page
		$TPREVPGLINK$	- Link to previous page of thread index
		$TPREVSUBJECT$	- Previous mesg subject in thread

	o  Removed resources:

		NOTSUBSORT, TSUBSORT

	o  Some changes to default resource settings.

	o  Reorganized code.  Some new libraries have been created
	   to help in maintenance.

	o  Source code has been put under SCCS revision control.


=======================================================================
07/12/96	(1.2.3)

	o  Extracted initialization of data structures into mhinit.pl.
	   The file is just required from the main source.

	o  Use q{} instead of qq{} when trying to read database file.
	   Should fix require problem under MS-DOS.

	o  Added <!--X-Reference-Id: ... --> comments at beginning
	   of messages.  May aid in database recovery techniques.

	o  ';'s are now deleted in filenames in mhexternal.pl
	   (applicable only when "usename" option specified)

	o  Added recognition of '/' when converted e-mail addresses
	   to mailto links in message headers.

	o  Simple fix to mhtxt2022.pl for execution under Perl 5.


=======================================================================
04/18/96	(1.2.2)

	o  Increased the speed performance of base64 decoding.  Speed
	   increase is much greater under Perl 4 than Perl 5.

	o  Added -time option to print out total CPU execution time.
	   Mainly used for debugging reasons (like checking on
	   base64 decoding times).  Time information is sent to
	   standard error.

	o  Added M2H_LOCKDELAY envariable and -lockdelay option.
	   Either can be used to adjust the sleep time between
	   attempts to lock the archive.

	o  Added -force option to override a lock on an archive if
	   attempts to lock fail.

	o  Added image/x-bmp and image/x-pcx to the default supported
	   MIME types.

	o  Ignore "Sv:" at the beginning of subjects when sorting by
	   subject.  "Sv:" is Danish for "Re:".

	o  Fixed bug in mhutil.pl where TIDXPGEND actually set
	   TIDXPGBEG.

	o  Dynamically define exclude_field routine after reading
	   user options.  exclude_field is utilized when formatting
	   a message header in HTML.  Defining the routine at run-time
	   helps reduce the regular expression overhead the old version
	   of the routine entailed.  There should be an increase in
	   overall execution time.


=======================================================================
03/22/96	(1.2.1)

	o  Added support for x-uuencode content-transfer-encoding.

	o  Added -locktries command-line option.

	o  Added the resource variable $OUTDIR$.

	o  mhexternal.pl filter will use the name parameter string
	   on the content-type field as the anchor text to the file
	   if there is no content-description.

	o  application/x-patch is recognized and processed by the
	   text/plain filter (mhtxtplain.pl).

	o  Fixed bug in install.me and osinit.pl where setting
	   $'PROG caused perl to terminate if $'DIRSEP was a backslash
	   (occured under MS-DOS usage).

	o  Fixed bug in install.me in the create_dir routine.
	   If $DIRSEP was a backslash, the regular expression setting
	   @a would cause perl to abort with an error.

	o  Fixed database bug where the MIMEARGS resource setting
	   was not being stored.

	o  Fixed index listing bug where a reverse listing was not
	   correct if an index size was specified less than the
	   current size of the archive.


=======================================================================
03/01/96	(1.2.0)

	o  Rewrote message parser routine so it will work under Perl 5
	   for multipart messages.  The rewrite also allows some
	   additional features that are mentioned below.

	o  The -mbox and -mh options are no longer required.  MHonArc
	   will automatically determine which mode to operate in
	   based upon the file arguments.  Hence, one can specify
	   MH folders and mailbox files on the same command-line.
	   Both options are ignored if specified.

	o  An HTML index of an archive contents can be generated to
	   standard output (-genidx).

	o  Message header lines not conforming to RFC 822 are ignored.
	   (Eg: Those pesky "From " lines should not show up anymore --
	    please do not confuse this with the regular "From:" lines;
	    note the colon vs the space).

	o  New resources:

		BOTLINKS	- May be used to completely customize
				  the links at the bottom of messages.
		IDXPGBEGIN	- Opening markup for main index page.
				  Allows one to redefine opening HTML
				  element, HEAD element, TITLE element,
				  opening BODY element, etc.
		IDXPGEND	- Closing markup for main index page.
		IDXSIZE		- Set the maximum number of messages
				  listed in index.  This is different
				  in MAXSIZE where MAXSIZE will remove
				  older messages when the MAXSIZE limit
				  is reached in the archive.
		MIMEARGS	- Define arguments to filters
		MSGPGBEGIN	- Opening markup for message pages.
				  Allows one to redefine opening HTML
				  element, HEAD element, TITLE element,
				  opening BODY element, etc.
		MSGPGEND	- Closing markup for message pages.
		NEXTBUTTON	- Defines the 'Next' button.
		NEXTBUTTONIA	- Defines the 'Next' button when it is
				  inactive.
		NEXTLINK	- Defines the 'Next' link.
		NEXTLINKIA	- Defines the 'Next' link when it is
				  inactive.
		NOTSUBSORT	- Do not sort threads by subject.
		OTHERINDEXES	- List other resource files defining
				  other indexes to create when creating,
				  or updating, an archive.
		PREVBUTTON	- Defines the 'Prev' button.
		PREVBUTTONIA	- Defines the 'Prev' button when it is
				  inactive.
		PREVLINK	- Defines the 'Prev' link.
		PREVLINKIA	- Defines the 'Prev' link when it is
				  inactive.
		TIDXPGBEGIN	- Opening markup for thread index page.
				  Allows one to redefine opening HTML
				  element, HEAD element, TITLE element,
				  opening BODY element, etc.
		TIDXPGEND	- Closing markup for thread index page.
		TOPLINKS	- May be used to completely customize
				  the buttons at the top of messages.
		TSUBSORT	- Sort threads listed by subject.

	o  Removed resources:

		INDEXBL, INDEXFL, MBOX, MH, NEXTBL, NEXTFL, PREVBL,
		PREVFL, TINDEXBL, TINDEXFL

	   Resource were removed because they were no longer applicable
	   and/or have been superceded by other resources.  MHonArc
	   will still honor old resource settings (where applicable) of
	   older archives and incorporate them into the new resource
	   settings.

	o  When specifying the resource file, mhonarc will now do
	   the following to determine its location:

		1.  If its an absolute pathname, mhonarc uses it.
		2.  If it is a relative pathname, mhonarc checks for
		    it relative to the current working directory.
		3.  Otherwise, mhonarc checks for it relative to
		    location of the archive as specified by outdir.

	   This resolution will allow you to place resource files with
	   the archive if desired (can be useful when using the
	   OTHERINDEXES resource element).

	o  Because of the new resources available, many <HR>'s are no
	   longer hard-coded and are controllable by resources.  <HR>'s
	   are still used in message pages to separate message data
	   from mhonarc data.

	o  Added resource variables:
	   (NOTE: Some variables are only valid in certain contexts)

		$DDMMYY$	- Date of message in dd/mm/yy format
		$IDXSIZE$	- Max size of index list
		$MMDDYY$	- Date of message in mm/dd/yy format
		$MSGID$		- Message id
		$NEXTBUTTON$	- Next button markup
		$NEXTFROM$	- From field of next listed message
		$NEXTFROMADDR$	- From e-mail address of next listed message
		$NEXTFROMNAME$	- From name of next listed message
		$NEXTLINK$	- Next link markup
		$NEXTMSGNUM$	- Number of next listed message
		$NEXTSUBJECT$	- Subject text of next listed message
		$NUMOFIDXMSG$	- Number of messages in index list
		$PREVBUTTON$	- Previous button markup
		$PREVFROM$	- From field of previous listed message
		$PREVFROMADDR$	- From e-mail address of prev listed message
		$PREVFROMNAME$	- From name of previous listed message
		$PREVLINK$	- Previous link markup
		$PREVMSGNUM$	- Number of previous message
		$PREVSUBJECT$	- Subject text of previous listed message
		$YYMMDD$	- Date of message in yy/mm/dd format

	o  Can specify a 'U' with variable length specifier to denote
	   replacement string is to be used in a URL.  Examples:

		$SUBJECTNA:40U$
		$MSGID:U$

	   The 'U' causes the replace text to have special characters
	   escaped as denoted by the URL spec.  NOTE: Specify ":U"
	   should NOT be used in the MAILTOURL resource; the variables
	   will automatically be expanded according to the URL spec.
	   Specifyind ":U" or a length specifier in the MAILTOURL
	   resource will prevent mhonarc from detecting the variable.

	o  New command-line options:

		-genidx		- Generate HTML index of archive contents
				  to stdout.
		-idxsize	- Maximum number of messages shown in indexes
		-notsubsort	- Do not sort threads listed by subject.
		-savemem	- Write message data while processing
		-tsubsort	- Sort threads listed by subject.

	o  The library mhtxt2022.pl has been added that provides
	   a filter to process ISO-2022 (Japanese) encoded mail messages.
	   See mhtxt2022.pl on how to hook it in.

	o  The mhexternal.pl filter by default ignores any filename
	   specification in the message for creating derived files.
	   This avoids name conflicts and security problems.
	   The "usename" filter option may be used to override this.

	o  Mime filters are now called with two additional arguments:

		$converted_data = &function(
			$header, *parsed_header_assoc_array,
			*message_data, $decoded_flag,
			$optional_filter_arguments);

	   The $decoded_flag is set to 1 if the *message_data has been
	   decoded.  $optional_filter_arguments contains an optional
	   argument string as determined by the filter.

	o  Mime filters can now be registered for multipart types
	   and message types.  This allows one to override mhonarc's
	   conversion of these types, and completely replace mhonarc's
	   message->HTML conversion process.

	o  Mime filters should now use $'FieldSep instead of $'X for
	   accessing parsed message headers.

	o  Mime filters can be registered for a base type.  Ie.
	   It is no longer required to explicitly list each possible
	   subtype if a single filter is to be used for them all.
	   Example:

		<mimefilters>
		image/*:myfilter'imagefilter:myfilter.pl
		</mimefilters>

	   Registers "myfilter'filter" for all image data types,
	   regardless of subtype.  However, if an explicit entry
	   exists for a subtype, then that filter is called.
	   Example:

		<mimefilters>
		image/*:myfilter'imagefilter:myfilter.pl
		image/gif:myfilter'giffilter:myfilter.pl
		</mimefilters>

	   "myfilter'giffilter" is called for all image/gif data.
	   "myfilter'imagefilter" is called for all other image data.

	o  A new resource, MIMEARGS, may be used to pass optional arguments
	   to filters to control their behavior.  The format of the
	   argument string is controlled by the various filters.  The
	   arguments can be specified by a specific content-type, or
	   for the filter routine in general.  A content-type argument
	   will be used over any arguments specified for a filter.

	   Example usage:

		<mimeargs>
		image/gif:inline usename
		m2h_external'filter:usename
		</mimeargs>

	   See the documentation for possible arguments to filters.

	o  Installation program can now be invoked in batch mode.

	o  Thread index properly includes docurl as main index.
	   -nodocurl will prevent the inclusion as with the main index.

	o  Fixed bug in mhtxthtml.pl on properly propogating a base
	   URL to relative URLs starting with a "/".

	o  Fixed bug where single quotes, and backslashed in keys of
	   associative arrays in the database file were not getting
	   escaped.

	o  Fixed bug where spaces and special characters were not
	   properly escaped in URL strings: spaces were left as-is,
	   and special characters were deleted.

	o  Removed illegal invocation choices in the Synopsis of the
	   documentation.


=======================================================================
04/24/95	(1.1.1)

	o  Fixed bug in -scan output where month in date was off by
	   one.


=======================================================================
04/21/95	(1.1.0)

	o  Made modifications to make MHonArc suitable to run
	   under MS-DOS without modification.  MHonArc will
	   automatically detect if it is running under Unix or MS-DOS.

	o  Added support for a thread index.  MHonArc will create a
	   complimentary index to the main index showing message
	   threads.

	o  Archive messages can be deleted.

	o  A listing to stdout of an archives contents can be
	   generated.

	o  Maximum number of messages for an archive can be set.
	   Older messages (based on sort method) are removed
	   automatically during add operations.

	o  MHonArc will now recognize if you try to add in a message
	   that already exists in an archive.

	o  The -editdx option will now also cause a updating of all
	   mail messages.  Guarantees resource changes to affect all
	   messages.

	o  Added the following resource file elements:

		MSGFOOT		-- Footer text for converted messages
		MSGHEAD		-- Header text for converted messages
		NODOC		-- Do not put link to documentation
		NOTHREAD	-- Do not create thread index
		TFOOT		-- Text at bottom of thread index page
		THEAD		-- Text at top of thread index page
		THREAD		-- Create thread index
		TLEVELS		-- Depth of thread listing
		TLITXT		-- Template text for entry in thread
				   index
		TIDXFNAME	-- Thread index filename
		TINDEXBL	-- Top button label in messages to
				   thread index
		TINDEXFL	-- Verbose label in message to thread
				   index
		TTITLE		-- Title of thread index page

	o  Added the following command-line options:

		-maxsize	-- Maxinum # messages in an archive
		-nodoc		-- Do not put link to documentation
		-nothread	-- Do not create thread index
		-rmm		-- Remove messages from an archive
		-scan		-- Listing of archive to stdout
		-thread		-- Create thread index
		-tidxfname	-- Thread index filename
		-tlevels	-- Depth of thread listing
		-ttitle		-- Title of thread index page

	o  Added the following environment variables:

		M2H_MAXSIZE	-- Maxinum # messages in an archive
		M2H_THREAD	-- If non-zero, create thread index
		M2H_TIDXFNAME	-- Thread index filename
		M2H_TLEVELS	-- Depth of thread listing
		M2H_TTITLE	-- Title of thread index page

	o  Added the following variables for template resources
	   (applicability of variables vary depending on the resource):

		$DOCURL$	-- URL to documentation
		$IDXFNAME$	-- Main index page filename
		$IDXTITLE$	-- Main index page title
		$NEXTMSG$	-- Next message filename
		$PREVMSG$	-- Previous message filename
		$PROG$		-- Program name
		$TIDXFNAME$	-- Thread index page filename
		$TIDXTITLE$	-- Thread index page title
		$VERSION$	-- Version number of the program

	o  Added $FROM$, $MSGID$, and $SUBJECT$ variables to be used
	   in the MAILTOURL resource.

	o  The string `$$' in template resources will produce a `$' in
	   the output.

	o  Fixed problem with messages (with follow-ups) getting
	   unnecessarily updated when messages are added to an archive.

	o  Only a CR/LF, or LF, pair will terminate a message head.
	   Before, MHonArc was terminating message heads when
	   encountering an empty line or a line that only contained
	   whitespace (which was incorrect behavior).

	o  Fixed bug in mhexternal.pl dealing with the `name' parameter
	   in the content-type field.  Surrounding "s or 's were not
	   being deleted causing filenames with quotes to be written.

	o  mhexternal.pl: The head of a pathname in the `name'
	   parameter in the content-type field is stripped off before
	   writing the external file.  I.e. Only the base filename is
	   used.

	o  Only one <HR> after the H1 subject in messages will appear
	   if no message header fields are printed.

	o  Added recognition of the following content-types in
	   mhexternal.pl:

		application/mac-binhex40

	o  Added a extras/ directory containing useful programs for
	   MHonArc.  See README in the directory for information on
	   the programs contained in there.

	o  To support -rmm, MIME filters now return an array.  The
	   first array value is the HTML for the message, and any other
	   array values are filenames of files generated by the
	   filter.  This allows MHonArc to know of any extra files that
	   must be deleted when a message is removed.

	o  Some routines from the main mhonarc source file have been
	   moved into a separate librarys: readmail.pl, mhdb.pl,
	   mhutil.pl

	o  The default URL to the documentation is now,

		http://www.oac.uci.edu/indiv/ehood/mhonarc.html

	   The old URL,

		http://www.oac.uci.edu/indiv/ehood/mhonarc.doc.html

	   is still valid.

	o  There's probably other stuff, but I cannot remember.


=======================================================================
10/01/94	(1.0.0)

	o  First release -- See RELNOTES about compatibility issues
	   with mail2html.


=======================================================================
#######################################################################
=======================================================================
Earl Hood, ehood@medusa.acs.uci.edu
@(#) CHANGES 97/06/06 19:47:06 @(#)