Revision History for txt2html
=============================

2.53 2019-08-02

  * New repository for this project (now in GitHub).
    https://github.com/resurrecting-open-source-projects/txt2html
  * Added CONTRIBUTING.md file.
  * Added README.md and renamed old README file to README.txt2html.
  * Fixed the licensing in LICENSE file.
  * Fixed some spelling errors.
  * Fixed the path for Perl in scripts/txt2html.
  * Unified the changelogs.



  |  -----------------------  |
  |  ---- OLD CHANGELOG ----  |
 \ / ---- OLD  UPSTREAM ---- \ /
  ·  -----------------------  ·

2.52.01 (aka 2.5201) 2013-05-21
-------------------------------

* 2013-05-21 16:25:10 +1000
  rebuilding with new version
* 2012-05-11 14:20:41 +1000
  Updating website.

2.51 Sun 4th March 2008
	- fixed bug with underscores in links
	- fixed docs about escape_chars (should be escapechars)
	- fixed docs about DOCTYPE

2.50 Sat 22nd December 2007
	- fixed bug with formatting and punctuation
	- removed old reference-to-an-array argument method
	- made --xhtml true by default (used to be false)
	- moved the debugging options to global variables

2.46 Fri 9th November 2007
	- updated docs on custom_heading_regexp
	- fixed bug with xhtml output
	- documented all undocumented functions

2.45 Fri 26th January 2007
	- fixed bug with umlauts
	- fixed bug with UTF-8 characters
	- added --underline_delimiter option.

2.44 Tue 17th January 2006
	- fixed bug with delimiter tables
	- minor documentation fixes

2.43 Fri 7th October 2005
	- fixed bug with interaction between #bolding# and #anchor links

2.42 Wed 10th August 2005
	- new option to txt2html script: --instring, which enables one
to process a string instead of a file.
	- new options to HTML::TextToHTML:
	  * instring, as above
	  * inhandle, which enables one to pass in input file handles to process
	  * outhandle, which enables one to pass in an output file handle

2.41 Sun 8th May 2005
	- solved the system links dictionary problem!  No longer uses
an external file at all, uses the DATA handle.  This means that there is
no longer a --system_link_dict option.
	- changed versioning scheme; now bugfix versions will be
things like 2.4101
	- removed the run_txt2html command; it isn't needed when the txt2html script is part of the package and does things much nicer.
	- generate the README from the PoD

2.40 Sat 12th February 2005
	- much improved speed

2.37 Thu 10th February 2005
	- another fix to installation
	- fixed CPAN module problem

2.36 Sun 23rd January 2005
	- slight fix to installation

2.35 Wed 18th January 2005
	- fixed bug where a Dos file was processed on a Unix system.
	- removed Makefile.PL; not needed with modern perls

2.34 Thu 6th January 2005
	- fixed another bug with demoronize code (gah!)

2.33 Wed 5th January 2005
	- darn, left out some files!

2.32 Wed 5th January 2005
	- fixed bug with demoronize code
	- fixed documentation for lower_case_tags
	- changed around installation so that no custom Module::Builder
is needed; also moved files into more customary places, and now use
Filter::Simple to change the system dictionary path in the module
and script files for installation.

2.31 Tue 21st September 2004
	- fixed bug with install
	- did some changes to help version-changing

2.30 Fri 27th August 2004
	- changed build system over to Module::Build to eliminate problems
with different versions of ExtUtils::MakeMaker and the installation of
the system links dictionary, as well as hoping to make the installation
more portable.
	- changed build so that the system links dictionary (in defaults
and documentation) is automatically set to where the system links
dictionary is going to be installed (yay!)
	- improved the INSTALL document (thanks to Mark Schmidt)
	- improved the bold and italic processing (now it's not messed up
by short lines putting a <BR> in)
	- bug fix for paragraphs with "0" (thanks to Andrew Williams)

2.25 Sun 23rd May 2004
	- the default location for the system links dictionary (txt2html.dict) is
now "/usr/local/share/txt2html/txt2html.dict". (It uses SITEPREFIX not PREFIX now)
	- added --bold_delimiter and --italic_delimiter options; now the bold and
italic matching is done separately, not as part of the links-dictionary stuff.
This means it's much easier to change or turn off if you don't want it.
	- links-dictionary matching is now done on segments without structural tags in them
so matching for things like '^<P>' should be replaced with just '^'. I don't think
this will break anything, and will prevent certain kinds of illegal substitutions.
	- tidied up some of the links in the links dictionary

2.24 Sun 16th May 2004
	- bug fixes for preformatting: fixed up loss of trailing PRE, and
now preformats and lists no longer have a trailing blank line
	- documented the 'quote_mail' and 'quote_explicit' classes, and added
the 'mail_header' class for mail-header paragraphs.
	- made the #bold# and *italic* matches more aggressive.
	- added Alan Jackson's "demoronize" patch from John Walker's demoronize script
at <http://www.fourmilab.ch/>

2.23 Wed 25th February 2004
	- oops! Forgot to include one of the test files!

2.22 Sun 22nd February 2004
	- bug fix with delimiter table

2.21 Sat 10th January 2004
	- bugfixes with a stricter perl 5.8.2
	- bugfix for processing empty file

2.20 Sun 7th December 2003
	- added --table_type option to say which types of table will
be recognised and parsed if the --make_tables option is true.  In other
words, there are now new types of table which can be parsed.
	    * ALIGN is the original space-aligned table type
	    * PGSQL is the type of table you get from a Postgresql query
	    * BORDER is a table with +-----+ lines around it as a border
	    * DELIM is delimited columns

2.10 Sat 6th December 2003
	- changed process_para method to assume it has only one paragraph;
and added process_chunk method to deal with processing strings which may
contain more than one paragraph in them.
	- a fair few internal changes to make more things think in terms of
paragraphs rather than lines; this changes the parsing of a few things,
but doesn't break the conventions.
	- fixed a bug where unordered lists were allowed to have bullets
which were more than one character wide.  Well, I considered it a bug,
it was annoying me.
	- added --bullets and --bullets_ordered options to enable the
user to define the bullet characters used for unordered and ordered
lists.
	- hooray!  Figured out a way of having multi-paragraph list items!
	- woo hoo!  Added definition lists!  And managed to do it so that
they are treated like the other lists, that is, you can nest lists in
definition lists and visa versa.

2.06 Sun 15th November 2003
	- fixed bug with processing STDIN (ie piping input to txt2html)
	- moved test input files into separate tfiles directory
	- added in most of Seth's old test files to the testing; and fixed
resulting bugs that were flushed out of cover

2.05 Wed 5th November 2003
	- error in fix to PREFIX! Argh!

2.04 Tue 4th November 2003
	- changed Makefile.PL to fix PREFIX
	- use "#!/usr/bin/env perl" trick (courtesy of Sami Haahtinen)
instead of using ExtUtils::configPL module
	- made Getopt::ArgvFile an official prerequisite
	- enabled CAPS tagging to be optional (just give an empty caps_tag)

2.03 Tue 15th July 2003
	- fixed bug with para tests (it didn't fail on my system
because it was using the already installed system links dictionary)

2.02 Sun 13th July 2003
	- fixed bug in documentation about custom headings
	- moved tests into t/ directory
	- added is_fragment option to process_para to enable it to
process a fragment without assuming that it was a paragraph.

2.01 Sun 1st June 2003
	- fixed up a few documentation/reference things that I'd forgotten,
changing names in sample.txt for example.

2.00 Sat 31st May 2003
	- merge of HTML::TextToHTML and the official txt2html - hence
the version jump.
	  * the distribution name is now txt2html
	  * renamed texthyper to txt2html, TextToHTML.dict to txt2html.dict
	  * merged change history
	  * split README into README, INSTALL and LICENSE files
	  * updated DEVNOTES
	- merged in the changes from 1.28 to 1.35 as much as I could
	  * corrected bugs that still applied
	  * added --style_url option
	  * added --body_deco option
	  * added two CSS classes: quote_mail and quote_explicit
	- Did Not add the following, as it was rather more complicated:
	  * multi-paragraph list items 
	  * the heading_callback_customize stuff

==================================================

1.12 Sat 15th February 2002
	- removed heavily spammed email address from documentation
and examples.

1.11 Fri 20th December 2002
	- fixed bug in texthyper script which was giving warnings.

1.10  Wed 18th December 2002
	- removed all dependency on AppConfig
	  * all of the existing options now must use their full names.
	  * However, one now has the choice between passing options
as a hash, or the old way, as a reference to an array.
	- removed the do_help method; if you want the documentation
of the module, use perldoc HTML::TextToHTML
	- moved the texthyper script into this distribution
	  * It now uses Getopt::Long and Getopt::ArgvFile.
	  * The format of .texthyperrc has changed to conform with
Getopt::ArgvFile rather than AppConfig.
	  * Changed the version number so that it was bigger than
either the script or the module so that both could have the same
version number (that's why the big jump).
	- the included system link dictionary (TextToHTML.dict)
is now installed in /usr/share/txt2html as part of the install process.

0.09  Wed 20th November 2002
	- improved the XHTML mode, so that open paragraphs get closed
sooner. This fixed a bug related to paragraphs inside lists.

0.08  Wed 20th November 2002
	- CPAN testers complained about a lack of explicitly stating
all the dependencies of AppConfig, which either means that AppConfig
has changed desperately, or their testing methods have changed, since
I didn't think it was possible to get the AppConfig module without getting
all its dependent modules, but, oh well.

0.07  Sun 17th November 2002
	- fixed a bug in process_para to ensure that if one is using it
standalone, any open lists will be closed
	- added --lower_case_tags option to force the tags to be output
in lower-case
	- first pass at XHTML conformance; added --xhtml option.
It isn't that pretty-looking, but the sample does pass the scrutiny of "tidy".
When turned on it:
	  * forces lower-case on
	  * makes empty tags have the empty marker (eg <hr/>)
	  * closes all open P and LI tags where they should be closed
	  * table cell alignments are done as style attributes

0.06  Mon 2nd September 2002
	- fine-tuned some of the links in the default links dictionary
	- some internal rewriting

0.05  Wed 5th June 2002
	- fixed minor bugs

0.04  Sun 2nd June 2002
	- fixed bug with detection of paragraphs by indentation
	- added --indent_par_break and --preserve_indent options
	- fixed error in documentation
	- fixed bug with nonexistant link dictionaries

0.03  Sun 26th May 2002
	- documented the format of the Link Dictionary
	- added the do_help method, and changed the behaviour of --help
and --manpage
	- added the --make_anchors option, which enables one to disable
the making of anchors, so that if one prefers another method of
anchor-making (such as that in HTML::GenToc) then one can use that
instead.
	- altered the #bold# pattern in the link dictionary to only need one
hash.  This should still hopefully allow things like #1 without turning it
bold, and being able to use ### as a separator.
	- gratuitous self-promotion: added HTML::GenToc to the
sample links dictionary
	- removed the need for getline(), but rather pass the lines in to
the methods, in order to parse by-paragraph and then by-line (mucho rewrite)
which enabled me to:
	  * implement the table-parsing from Gareth Rees's HTML::FromText
module -- added the --make_tables option
	  * now the links dictionary does multiline matches (useful for
things like italics which break over lines)
	  * enable converting passed-in strings rather than just files

0.02  Wed 15th May 2002
	- fixed bug with link dictionary parsing
	- improved the tests
	- updated link dictionary to fix a few bugs (eg underlines)
and add a few things (like using double # for ##bold## text).

0.01  Sun 12th May 2002
	- conversion of Seth Golub's txt2html (version 1.28) to a module
	- made all global settings options (eg the location of the system
link dictionary)
	- added "outfile" option
	- added use_mosaic_header option
	- changed the dynamic code generation completely.
	- removed the evil $* variable

-----------------------------------------------------------------

=====================================================
(cvs log on Sourceforge)

revision 1.35
date: 2002/12/03 03:47:46;  author: suntong;  state: Exp;  lines: +14 -13
- Defines two CSS Style: quote_mail & quote_explicit
- Mail quote mode and quote mode won't interfere with each other
----------------------------
revision 1.34
date: 2002/12/03 01:51:41;  author: suntong;  state: Exp;  lines: +8 -7
- preformat should be dealed before other cases
----------------------------
revision 1.33
date: 2002/12/02 23:44:13;  author: suntong;  state: Exp;  lines: +7 -7
Bug fixing (details in http://groups.yahoo.com/group/txt2html/message/89):
- Changed incorrect entity names of "fraqfrac*" to "frac*".
- redundant \| in [\||:] removed.
- change misleading --style help to '--style <URL>'
----------------------------
revision 1.32
date: 2002/12/01 19:59:09;  author: suntong;  state: Exp;  lines: +39 -5
- Add Callback functions for HTML header handling
  so that users can customise their own heading,
  add horizon lines, change colors or write their own toc, etc

- User can define/keep their callbacks locally without tampering main
  distribution

- 'mailstuff' priority should be higher than 'preformat'
----------------------------
revision 1.31
date: 2002/12/01 18:59:30;  author: suntong;  state: Exp;  lines: +5 -5
- solve the problem that txt2html can't html-ify its own anchor.
----------------------------
revision 1.30
date: 2002/12/01 18:51:20;  author: suntong;  state: Exp;  lines: +53 -10
- All my previous patches are lost because yahoogroups doesn't keep my patch
  attachements. So, here they are in a big chunk. Major updates are:
- Able to use a style sheet for the generated html file
- Able to use body decoration for the generated html file
        customize your own background color/image/sound, etc...
- Misc enhancements.
----------------------------
revision 1.29
date: 2002/12/01 18:28:35;  author: suntong;  state: Exp;  lines: +3 -2
- Apply patch from http://groups.yahoo.com/group/txt2html/message/32

,-----
| I don't know when I'll get around to releasing a new version of
| txt2html, but I have a few fixes I've been sitting on. I thought
| I'd send them out on this list so people could take advantage of
| them without having to wait until I finally package up a new release.
|
| * Changed incorrect entity names of "fraq*" to "frac*".
|
| * Allow paragraphs within lists by permitting blank lines within
| list elements, as long as the following text has the same
| indentation level.
`-----

=======================================================

1.28
----

- bugfix: reserved characters in titles created with --titlefirst are
  now escaped properly.

- bugfix: when preformatting entire document, each line was
  getting its own <PRE></PRE> container (introduced
  with explicit preformatting feature in 1.26).

- dict: added some characters to those allowed in http urls (=&;,).

- dict: added "-" to allowed characters within *emphasized-pattern*.

1.27
----

- Changed names of default link dictionaries to txt2html.dict


1.26 (not released)
-------------------

- Added -8 (for 8-bit-clean) to disable conversion of non-ASCII
  characters to their corresponding Latin-1 character entities.

- Added -pm to allow explicit marking of preformatted text in source

- Changes => to , in mapping, to stay compatible with Perl 4

- Added debug flag 4, for observing link rules in action

- Fixed length checking bug in header underline analysis

- Change a regexp so Perl 5.6 doesn't complain.

- No longer add space after <LI> tags

- Allow unindented lists to start after CAPS lines

- Use · as a bullet character

- Fixed bug that dropped a character when certain actions were
  taken on the last line of input that didn't end with a newline.

- Added more aggressive regexps for _underlined_ and *emphasized* text.

- Improved character markup rules

- Added link rule for news URLs.  (This must have been
  accidentally deleted at some point.)

- Added link rule for common explicit url markup: <URL:foo>

1.25
----

- Changed the official home page to <http://www.aigeek.com/txt2html/>
  (the old page will have a working redirect indefinitely.)

- Added a LICENSE to the distribution.  (modified BSD-style)

- When no title is specified, an empty title element is inserted.
  (The old behavior was to omit the title element, which is
  forbidden by the spec.)

- Made heading anchors appear inside the heading, rather than
  surrounding it (which is forbidden by the HTML spec)

- Changed the DTD name

- Added the --linkonly option so people can use the links
  dictionary feature without doing any other markup.  This is
  useful for adding links to HTML fragments or documents.

- Added the --prepend_body option for prepending HTML to the body.

- Made in_link_context smarter so it won't link on attributes or
  tag names.  (This is good for adding hyperlinks, but may screw
  up some clever uses of the linking code.)

- Added link rules for _underlined text_ and *emphasized text*

- Added --noescapechars to suppress converting "&" "<" and
  ">" into "&amp;" "&lt;" and "&gt;"

- Changed pattern rules to handle non-ascii letters properly in
  matching patterns.

- Added conversion of non-ascii letters into character entities.

- Lots of upgrades to the links dictionary patterns

1.24
----

- Changed behavior of custom headers to something much more
  useful: Header levels are assigned by regex in order seen.
  When a line matches a custom header regex, it is tagged as a
  header.  If it's the first time that particular regex has
  matched, the next available header level is associated with it
  and applied to the line.  Any later matches of that regex will
  use the same header level.

- Added the -EH / --explicit-headings option

- Added some unnecessary initialization to avoid warnings when
  perl is run with the -w switch.

1.23
----

- Added handling for when the consistent formatting of numbered
  lists is the position of the non-numeric character, not the
  amount of whitespace preceding the number.  (The numbers
  grow to the left instead of the right.)

1.22
----

- Fixed bug in unhyphenation

- Changed HTML version in default doctype line to 3.2

1.21
----

- Added <META NAME="generator" CONTENT="txt2html v1.21">

1.20
----

- Added DOCTYPE tag and --doctype options.

- Syntax change to get rid of Perl 5 warning

- Added ability to use the first line of the text as the title

- Fixed some (unused) grossness in links dict file

1.19
----

- Added --append_head

- Mail and News name anchor surrounds just the first word
  ("Newsgroups:" or "From"), and not the whole line.  That way,
  newsgroup names and email addresses get HREF'd as normal.

1.18
----

- Cleaned up nested list handling & fixed a bug under Perl 5.

- Changed a couple minor things to get rid of some of the Perl 5 warnings.

1.17
----

- Lists can start even when not indented and not preceded by a
  blank line if the previous line was short or a header.

- New flag "o" added for dictionary entries.  Specifies that the
  link should only be done the first time a match is found.

1.16
----

- Added anchoring of custom headers

- Took the changelog out of the script

- Tweaked $line_indent in sub liststuff

- Insert <P> before each mail/news message

1.15
----

- Fixed options handling for -e/+e , -r

- Added "Newsgroups:" to trigger mail headers

- Fixed anchor naming

- took out -T option, since it isn't implemented yet.  Whoops..

- Fixed bug in endpreformat

1.14
----

- Fixed +l/--nolink option handling

- Fixed major bug in dynamic_make_dictionary_links that allowed
  nested links under some circumstances.

1.13
----

- Fixed usage message so it matches options.  (whoops)

- Added custom heading style feature

1.12
----

- Fixed bug in heading regexp

- Changed underline tolerance parameters from min & max length
  difference to length difference & offset difference

- Centralized line reading, added handling of DOS carriage returns

- Switched to heading style stack.  Styles still very limited.

- Changed heading anchor names from a simple count to a hierarchical
  section number.

1.11
----

- Blank lines are never considered underlined

- Shortline breaking slightly more intelligent (or at least different)

- Paragraph breaks much more intelligent

- Lowercased tags.  Style is so fickle.

- Added links dictionaries, link making, etc.

- Allow repeated bullet chars for unordered lists.  (Tiny mod to regexp)

- switched order of caps & liststuff in main()

- improved untabify() so it converts the whole line, not just beginning

- split up all lines >79 characters to avoid common downloading error
  (people would sometimes copy the script off the display,
  inadvertently adding a few newlines in bad places in the code)

- Handles option "--" now.

- Accepts named files as input as alternative to stdin

- Deals with stdin properly (no more extra EOFs needed)

- Improved mail handling

1.10
====

- Added --extract, etc.

1.9
---

- Changed from #!/usr/local/bin/perl to the more clever version in
  the man page.  (How did I manage not to read this for so long?)

- Swapped hrule & header back to handle double lines.  Why should
  this order screw up headers?


1.8
---

- put mail_anchor back in.  (Why did I take this out?)

- Finally added handling of lettered lists (ordered lists marked with
  letters)

- Added title option (--title, -t)

- Shortline now looks at how long the line was before txt2html
  started adding tags.   ($line_length)

- Changed list references to scalars where appropriate. (@foo[0] -> $foo[0])

- Added untabify() to homogenize leading indentation for list
  prefixes and functions that use line length

- Added "underline tolerance" for when underlines are not exactly the
  same length as what they underline.

- Added error message for unrecognized options

- removed \w matching on --capstag

- Tagline now removes leading & trailing whitespace before tagging

- swapped order of caps & heading in main loop

- Cleaned up code for speed and to get rid of warnings

- Added more restrictions to something being a mail header

- Added indentation for lists, just to make the output more readable.

- Fixed major bug in lists: $OL and $UL were never set, so when a
  list was ended "</UL>" was *always* used!

- swapped order of hrule & header to properly handle long underlines


1.7
---

- Added to comments in options section

- renamed blank to is_blank

- Page break is converted to horizontal rule <HR>

- moved usage subroutine up top so people who look through code see
  it sooner


1.6
---

- Creates anchors at each heading

1.5
---

- Fixed minor bug in Headers

- Preformatting can be set to only start/stop when TWO lines of
  [non]formatted-looking-text are encountered.  Old behavior is still
  possible through command line options (-pb 1 -pe 1).

- Can preformat entire document (-pb 0) or disable preformatting
  completely (-pe 0).

- Fixed minor bug in CAPS handling (paragraph breaks broke)

- Puts paragraph tags *before* paragraphs, not just between them.


1.4
---

- Allow ':' for numbered lists (e.g. "1: Figs")

- Whitespace at end of line will not start or end preformatting

- Mailmode is now off by default

- Doesn't break short lines if they are the first line in a list
  item.  It *should* break them anyway if the next line is a
  continuation of the list item, but I haven't dealt with this yet.

- Added action on lines that are all capital letters.  You can change
  how these lines get tagged, as well as the minimum number of
  consecutive capital letters required to fire off this action.


1.3
---

- Tiny bugfix in unhyphenation


1.2
---

- Added unhyphenation

<http://www.aigeek.com/> seth@aigeek.com