# LaTeX2HTML by Nikos Drakos
# **************************************************
# Change Log ***************************************
# **************************************************
#
# Please use the following style: Add your (unique) initials to the list below.
# Use your initials whereever you make commented changes.
# To get in correlation with the more verbous explanations you make in the
# change log, or if you want to stress meaningful changes, use an additional
# short (about 3 letters) identifier, unique together with your initials,
# eg. jcl-pag. This also helps others to find your changes.
# If any change affects former changes, this *must* be stated in the log,
# together with a reference to the previous log entrie(s).
#
# nd = Nikos Drakos
# ab = Axel Belinfante
# ac = Andrew Cole
# wb = William Perry
# bh = Bob Hendley
# rst = Robert S. Thau
# dm = David Martland
# rjw = Roderick Williams
# jr = Jussi Rahola
# ec = Eric Carroll
# tl =
# fz = Franz Vojik
# spqr= Sebastian Rahtz
# up = Ulrich Pfeifer
# aa = Alberto Accomazzi
# wr = Wolfgang Riedel
# alo= Alex Lopez-Ortiz
# mb = Michael Francis Brantley
# mcb= Michael C. Grant
# jz = Jelle van Zeijl
# bs =
# jhf = James Freeman
# hpr = Hal Peterson
# kc = Ken Cox
# gp = George Phillips
# tw = Thomas Wieland
# dh = Dirk Husemann
# bt = Brian Toonen
# gg = Graham Gough
# kz = Kenny Zalewski
# tm = Tim MacKenzie (t.mackenzie@trl.oz.au)
# mk = Martijn Koster (m.koster@nexor.co.uk)
# ew = Evan Welsh
# re = Robert Estes
# rw = Richard Walker
# sl = Stefan Lohrum
# kl = Kris Laporte
# gg = Gordon Greene
# pb = Pascal Brisset
# po = Pieter Olivier
# eb = Ethan Bradford
# ev = Eric Verbeek
# tt = Tom Tromey
# ms = Mike Stok
# sk = Stefan Kaes
# psm = Peter Magnusson
# mb = Martin Boyer
#segal= Mark Segal
# jmn = Mark Noworolski
# meh = Marcus Hennecke
# hws = Herbert W Swan
# jkr = Jens Krinke
# sid = Sidik Isani
# rml = Rob Malouf
# jtc = Ted Cox
# pns = Paulo Ney de Souza
# tkm = Tom Miller
# mg = Michel Goossens
# ays = Andreas Schott
# jfr = J. Freeman
# mer = Michael Ernst
# rrm = Ross Moore
# jcl = Jens Lippmann
# dtp = D. Taupin
# mwk = Martin Wilck
# mcm = Barry McMullin
# tak = Takashi Tomokiyo
# mro = Marek Rouchal
# bv = Boris Veytsman
# kr = Keith Refson
# uw = Uli Wortmann
#
#
#------------------- Test Suite Manifest ----------------------------------
#
# A listing of current test suites.
# These suites should be run before any new version is released.
# For a detailed description consult the suite's header.
# Feel free to devise further test suites, but *comment* them (see
# jcl-pag.tex header on how to comment).
#
# docs/manual.tex The l2h manual.
# example/report.tex An example with segment document parts.
# tests/pages.tex Tests image generation of l2h.
# tests/rrm-col.tex See log.
# tests/jcl-pag.tex See log.
# tests/jcl-env.tex See log.
# tests/jcl-verb.tex See log.
#
#--------------------------------------------------------------------------
#
# CHANGE LOG FOR THE LATEX2HTML RELEASE
#
# What is recorded here are changes to any file of the LaTeX2HTML release
# *meaningful to the user*, such as new options or files, important bug
# fixes etc.
# A more detailed log history can be found within the particular files, near
# the top.
#
#
#--------- v98.1 (compiled by jcl)
#
# Changes in the top-level directory:
#
# rrm latex2html.config
# jcl - New section header strings in &english_titles (corresponding
# entries also in german.perl, french.perl):
# $ref_title - this text heads the References section
# $pre_title - for preface
# $child_name - this text heads the table of child links
# $prf_name - for proofs
# - added variables for tex and initex for use with the -ldump
# feature
# The -ldump switch will help to speed up image generation
# with *subsequent* runs on your document.
# - added command line option -unsegment and $UNSEGMENT
# Use latex2html -unsegment, or texexpand -unsegment, or set
# $UNSEGMENT to 1 in latex2html.config
# (see texexpand changes)
# - added defaults for bolder images of gray background
# - provide optional value for $TMP --- location of a tmp directory
# - provide default for $GENERIC_WORDS to be omitted from filenames
# when using the -long_titles switch
# (see also german.perl, french.perl)
#
# rrm latex2html
# (uw) - patch for -no_subdir bug, reported by Thomas Anders dta@dsapp2.hmi.de
# (jcl) - &cleanup deletes ${PREFIX}images.dvi
# - fixed bug whereby \item s were combined --- lack of unique brace id
# - fixed missing-space bug that killed math-images
# - implemented scheme to allow alternative address formats
# - address-data default is ISO conformant
# - added translation of \institute as part of the title page settings
# - change to -link so that -link + puts mini-TOC on leaf pages
# and -link puts no mini-TOC on these lowest-level pages;
# Alternatively, set $MAX_LINK_DEPTH to be negative for these mini-TOCs
# - new switch -ldump to use ldump.sty when making images,
# this speeds up processing on the 2nd and subsequent runs;
# - fully implemented dependency of counters;
# - implemented \expandafter and \noexpand with special care concerning
# case-changing macros;
# - fixed bug whereby `0' in math was getting lost;
# - allow alternative source for icons, using $ALTERNATIVE_ICONS
# - implemented on-the-fly \renewcommand
# - \renewcommand meta-commands are now wrapped, so that they are
# re-processed *before* there replacement text is processed.
# Otherwise bad LaTeX code can get into images.tex
# (Thanks to Michael Hall for an example of this.)
# - \newcommand both warns *and* replaces existing macros
# - more care over \limits macro
# - fixed a bug whereby some macro-expansions didn't work correctly
# - fixed printing of \the macros, with dependent counters
# - NEW: implementation of document-classes and styles
# allows file for classes and options and option-class pairs
# - implemented $INFO-page options
# - fixed error with minipage footnote numbering
# - implemented \numberwithin (from AMS-packages)
# - fixed error in DTD declaration
# - implemented \nobreakspace and \nonbreakingspace (AMS-packages)
# - Added support for Override.pm to latex2html
# - added command line option -unsegment and $UNSEGMENT
# Use latex2html -unsegment, or texexpand -unsegment, or set $UNSEGMENT
# to 1 in latex2html.config (see texexpand Changes)
# - **major**
# introduced a font/size state-stack concept to handle nesting
# of tags altering size and font-face within text-level data.
# It also affects macros which create block-level tags.
# The @open-tags list has to be cleared, tags inserted,
# then @open-tags recovered for the new paragraph, say.
# - font/size-changing commands are declared as `deferred'
# so as to be processed in sequence with environments
# \par \bigskip \medskip etc. are similarly `deferred'
# - the switch -tmp 0 uses /tmp as a tmp-directory
# - special characters can be `escaped' in macros that construct
# {rawhtml} environments
# - \newcommand \newenvironment \def definitions are more efficient
# macro substitution is *much* faster, using an array mechanism
# - support for HTML 4.0, including style-sheets
# - faster processing on user-defined macro replacements
# - faster processing of .aux files
# - improvments to ensure validity of the HTML code, wrt DTDs
# - many LaTeX macros use an extra indirection, so that it
# becomes much easier to cope with tag-balancing and styles;
# this also shortens the code somewhat.
# - the declared environments are similarly handled indirectly.
# - processing of the .aux file and user-defined macros handle
# smaller chunks of text, rather than the whole document;
# for the .aux file, each line is handled separately
# for macro-replacement, a \par command serves as a delimiter
# (since it is illegal within most (La)TeX macros).
# - other parts of the processing are done using arrays, to lessen
# the length of strings that might otherwise be copied.
# - when $USING_STYLES is set, a CSS style-sheet is generated,
# indicating the style used in each environment, and providing
# a unique ID to such environments. This CSS style-sheet can be
# subsequently edited, if desired.
# - \label and \index commands in section-headings work better,
# (the anchor is placed before the heading tag)
# also a forward-search is now conducted for these.
# - handling of charsets and input encodings is much improved.
# - moved some variables to establish defaults before .config or -init
# files are read.
# - introduced variables $image_pre and $IMAGE_PREFIX to control names
# used for the intermediate files created by dvips and pstoimg
# - implemented \textsc fully --- now 8-bit characters are case-folded too
# - create images anti-aliased against a gray background. Characters are
# bolder, without (almost) white edges against gray.
# new variables $LOAD_LATEX_COLOR and $LATEX_COLOR control this.
# - new switch -numbered_footnotes setting $NUMBERED_FOOTNOTES
# - \begin/endgroup and $b/egroup commands write directly into images.tex
# - defined &do_cmd_footnote* since \footnote*{...} is a common in TeX
# - defined &do_cmd_MF for the Metafont logo
# - support for pre-processors when making images of non-latin script
# code for generating images is placed into file: images.pre
# pre-processor commands are placed into file: preproc
# preproc is run as a script, producing: images.tex
# image generation then proceeds as normal.
# - pseudo-environments for font/styles are given better support
# and TeX's \font command is implemented
# e.g. {\bf ... } {\wncyr .... }
# the control-sequence is found and the grouping is treated as an
# environment, rather than waiting to handle it as a command.
# - theorem commands and styles are implemented;
# covers the {theorem} and {amsthm} packages.
# - footnotes have hyperlink back into the main text.
# - footnotes can have images in the footnote marker `word'
# - section numbering is rationalised to be more in-step with LaTeX
# the HTML numbering is unchanged, via @curr_sec_id
# in-text macros, like \value \setcounter \thesection get their
# number from $global{'section'} etc.
# \section advances both counters, but \section* only the HTML one.
# - changed the $NO_ACCENT_IMAGES variable instance to $ACCENT_IMAGES
# - added %numbered_section hash for initialising $global counters
# in &initialise_sections
# - some name changes of auxiliary files usage of temporary directories
# to get it running on DOS (thanks to Daniel Taupin)
# - implemented \theoremstyle , \theorembodyfont , etc.
# - added to the TeX code in images.tex for finer control with accents
# - inserted binmode(IMAGE) when scanning bitmaps for size
# (thanks, Daniel Taupin)
# - changes to the way captions are found and handled
# - new switch: -iso_language to get the correct DTD with mixed-languages
# - minipage environments can capture a caption, when inside figure/table
# - implemented \eqref (from AMS-TeX)
# - implemented \topcaption \bottomcaption \middlecaption commands
# - implemented use of 2 temp directories:
# $TEMPDIR for image-processing, usually subdir of /tmp
# TMP subdir of working directory, for the (formerly) TMP_ files
# - sectioning commands in expansions of \newcommand work properly now
# (thanks to Azat Badretdinov for finding the cause)
# - improved the handling of simple math generally
# - changed name \HTML to \HTMLcode ; \HTML remains as an alias
# - minor changes in images.tex to avoid some `cropping bars' remaining
# - itemize works properly, with explicit \item[...]
# bullets implemented as images, for other \item s
# - \tag is recognised in equation environments, with all math modes
# - \authorURL make the \author information into a hyperlink
# - \email adds to the title-page info
# - made \large and \small use tags
# - added declarations for \smaller and \larger
# - change to &do_cmd_mbox so that it doesn't try to make an image
# of the code that may already have an image_marker
# This allows \mbox{$...$} to be used with macro-replacements
# - improvements to &make_long_title to cope with accents, etc.
# - implemented -no_fork switch for $NOFORK variable
# - replaced system `cat` command with a platform-independent version
# - implemented \usecounter use within {list} environments
# - long-titles are restricted to 27 characters +(.html) = 32 chars
# - suppressed warning messages for `array` and `theorem` packages
# - use $BODYTEXT with the footnotes page (thanks Andreas Otte)
# - figure/table captions aren't also included in images
#
# NEW Override.pm
# This file is intended to hold OS-specific definitions that
# the latex2html or other scripts may require during runtime.
# Thanks go to Marcus Hennecke, Axel Ramge, Uli Wortmann and
# many others who helped with their suggestions.
# This is a step towards an operating system independent
# LaTeX2HTML which is currently under work.
#
# jcl configure-pstoimg
# - safer renaming of substitute files to .bak (lost files reported)
# - dies with message unless initialized via install-test
#
# jcl install-test
# rrm - some revisions for 98.1, see log of that file
#
# rrm pstoimg, pstoimg_nopipes
# uw - adjusted to use the Override.pm module with Perl5
# jcl - removed the PaperSize bug
# - handles absolute file names and temp dirs correctly
#
# jcl texexpand
# rrm - small fix to &interprete, \input|include now doesn't loose the
# uw comment if merging fails
# - added command line option -unsegment and $UNSEGMENT
# Use latex2html -unsegment, or texexpand -unsegment, or set
# $UNSEGMENT to 1 in latex2html.config.
# This will force a segmented document to expand its segment
# files, so that it may be processed as a whole with LaTeX2HTML.
# Use this feature to test a segmented document or whenever a
# document needs to be fully expanded.
# - Added support for Override.pm to texepand.
# - generalised pattern for matching verbatim-like environments
#
# NEW packages:
#
# IndicTeX-HTML/
# rrm provided IndicTeX-HTML, a collection of Perl and LaTeX styles
# to translate IndicTeX documents (see README)
#
# foilhtml/
# bv provided FoilHTML, a package for translating FoilTeX documents
# (see readme.v12)
#
# Changes in the styles/ directory:
#
# rrm alltt.perl
# reimplemented the {alltt} to use rather than
now
# style changes can be handled properly, with valid tag nesting
#
# NEW rrm amsart.perl, amsbook.perl
# New files --- support AMS document class
#
# rrm amsfonts.perl
# pass \mathbb commands to LaTeX
#
# rrm amsmath.perl, amstex.perl
# various changes (see log history in that files)
#
# rrm babel.perl
# - parametrised the loading-mechanism:
# a single line now suffices for each new language
# - added such lines for `french' and `francais' (NB. Michel Goossens)
# -- fixed error in file-loading
#
# NEW rrm article.perl, book.perl, letter.perl, report.perl, slides.perl
# Document-class emulation file
# -- adjusts the \the macros for sectioning commands
# -- suppresses warnings for standard class-options
#
# NEW kr chemsym.perl
# There may be some problems with accents, whose TeX macro names
# correspond to chemical symbols, as defined here --- there is no
# compensation provided within this implementation.
#
# rrm latex2html.config, francais.perl, french.perl
# jcl german.perl, germanb.perl
# - with -long_titles, use $GENERIC_WORDS list to shorten the title.
# - brought &german/french_titles up to date with &english_titles
#
# rrm graphics.perl, graphicx.perl
# - modified \graphicspath to parse a list of directories (Axel Ramge)
# - rewritten the &do_cmd_graphicspath subroutine
# works correctly and allows multiple use, outside the preamble
# - handle \DeclareGraphicsRule and \DeclareGraphicsExtension
# - \graphicspath works correctly in the preamble now
#
# rrm heqn.perl
# recognise \htmlborder and \htmlimage also in processed environments
#
# rrm html.perl
# \htmlmeta defined, for inserting tags
#
# NEW rrm inputenc.perl
# Implements the inputenc package, allowing the character encoding
# to be specified from within the document.
# Supports latin1, latin2, latin3, latin4, latin5, latin6, Unicode.
# This uses files latin1.pl to latin6.pl and unicode.pl
# in the versions/ directory.
#
# rrm makeidx.perl
# -- fixed problem with sub-item keys printing multiply
# -- \index commands within section-heads now work correctly
#
# NEW rrm more_amsmath.perl
# That part of amstex.perl and amsmath.perl that needs the `math'
# extension has been split-off into more_amsmath.perl .
# This is loaded automatically with switches:
# -no_math -html_version ...,math
#
# rrm natbib.perl
# Updated for compatibility with natbib.sty v6.6
# - all \cite... commands have a *-version and 2 optional arguments
# - Harvard emulation is now automatic
# - implemented \citep*
# - fixed \harvardurl to work properly and without html.sty
# thanks to James A. Bednar for noticing
#
# NEW rrm seminar.perl
# implements seminar.sty by:
# making \newslide into a sectioning-command, with its own numbering
# other commands are recognised and largely ignored
# text of `notes' are gobbled
#
# rrm supertabular.perl
# implemented the \tablehead \tabletail etc. commands
#
# rrm xy.perl
# - allow for optional argument to \newxycolor and \newgraphescape
# - recognise \htmlborder and \htmlimage also in processed environments
#
# Changes in the versions/ directory:
#
# jcl html2_1.pl, html2_2.pl, html3_0.pl, html3_1.pl, html3_2.pl
# renamed html*.*.pl files, for DOS
#
# rrm html2_2.pl (equal to table.pl)
# {tabular} environments could pick up caption from previous {table}
#
# rrm html3_1.pl (equal to math.pl)
# with $NO_SIMPLE_MATH, force images of \...matrix environments
# - fixed bug which affected the code for image-making
# - fixed bug not handling \lefteqn in {eqnarray}s
# - allow Netscape's VALIGN="BASELINE" for {eqnarray} rows
# - correctly writes {eqnarray*}, not {eqnarraystar} to images.tex
# - recognises \le (as well as \leq) without warning (thanks Michel)
# - automatic recognition of the differential `d' after \int
# - significant additions to math-parsing
# additions:
# \mathop,\mathrel,\mathbin, etc. recognised;
# sup/subscripts handled a little differently;
# array-like environments need to have delimiters protected,
# when they occur as sub-environments of other array-like envs;
# get equation-number from \\theequation not &do_cmd_....
# equation-numbers can have variable delimiters, not just (..);
# some AMS commands implemented always
# \text trated as \mbox --- not best, only temporary
# - fixed spacing/line-break problems
# - implemented \text properly, within math
# - fixed `missing 0' error with sup/subscripts
# - adapted for $USING_STYLES ; e.g. with HTML 4.0
# - adapted to allow use of Unicode entities for math-symbols
# - allow some fractions to be given as named entities; e.g. frac12
# - font styles/faces are now supported with properly nested tags
# - support for \strut and \vrule and \hrule
# - use the MATH and BOLDMATH classes when $USING_STYLES
# - always make images of some math operators
# - always make image of variable-sized operators
# - catch fractions via \frac \tfrac \dfrac
# - use a specified WIDTH="10%" for equation-numbering cells
# - made all numbering environments have WIDTH="100%"
# - introduced variables $math_start_rx and $math_end_rx .
# (looking ahead to the possibility of allowing $ to be non-math.)
# - fixed problem with superscripts: ' (\prime) missing from images
# - \mathbf , \mathcal etc. seem to be correct now
# - fixed bug causing some fractions to be in \textstyle, not displayed
# - fixed some string tests that incorrectly treated 0 as ''
# - recognise \htmlimage and \htmlborder in processed environments also
# ... and various other changes (see log history of that file)
#
# rrm html3_2.pl
# - Implemented {tabular*} environment, to be same as {tabular}
# i.e. does not recognise a fixed width
# - the \extracolsep macro is recognised, but ignored
# - sub-environments of {tabular}s are processed first;
# this allows {tabular}s to be nested
# - \multicolumn now implemented
# - cells containing `0' now show it, rather than being empty
# - {eqnarray} math-style bug fixed
# - allow Netscape's VALIGN="BASELINE" for {eqnarray} rows
# - removed unnecessary tags, when {eqnarray} cell is an image
# - reworked the handling of @-expressions in tabular specs.
# - remove &simple_math_env acting twice on the same math --- Ugh!
# - implemented the extra tabular declarations from array.sty
# - ignore and \PreserveBackslash commands
# these extensions are **untested**
# - adapted the alignment environments to work as instances of
# &declared_env which respects the font/size state-stack.
# - adapted some environment-like commands to work with style-sheets
# - fixed migrating captions error; thanks to Keith Andrews
# - use a specified WIDTH="10%" for equation-numbering cells
# - made all numbering environments have WIDTH="100%"
# - improvements to {eqnarray} environments:
# each field is handled separately, equation-numbering works
# eqn-numbering aligned OK now;
# - VALIGN on rows in {tabular} is not used, unless $NETSCAPE_HTML
# - introduced \mathon and \mathoff for use in tabular-specs
# - implemented \tablehead and \tabletail (from supertabular.sty)
# - implemented the @{...} and !{...} column-specs as separate columns
# without slots for filling from the table-data.
# - \tag and \notag now work correctly in normal math-mode
# - fixed some string tests that incorrectly treated 0 as ''
# - recognise \htmlimage and \htmlborder in processed environments also
# - make sure math within do_math_env is properly delimited
# - p{} in a {tabular}'s col-spec causes VALIGN="TOP"
#
# NEW rrm html4_0.pl
# Implements output specific for HTML 4.0
# This requires LaTeX2HTML v97.2 (alpha) or later version.
# The variable $USING_STYLES is set, to enable automatic creation
# of a CSS style-sheet.
# To use a UTF-7 or UTF-8 Unicode charset, the unicode extension
# needs to be loaded also; else the Latin-1 charset is used.
#
# NEW rrm lang.pl
# contains the language codes, from i18n.pl
#
# Changes in the versions/ directory:
#
# rrm html.sty
# - uses Robin Fairbairns' code for ignored environments,
# replacing the previous comment.sty stuff.
# - extensions to the \tableofchildlinks command
# - extensions to the \htmlinfo command
# - allow the dummy {tex2html_nowrap} environment in LaTeX
# use it to make its contents be evaluated in environment order
# - implemented an optional argument to \begin for style-sheet info.
# - modified use of an optional argument with sectioning-commands
#
# NEW rrm ldump.sty
# ldump.sty --- LaTeX package which allows format dumps to be used.
# use the new -ldump switch to make use of this feature/package.
#
#
#
#--------- v97.1 (compiled mostly by rrm and also jcl)
# incorporating the changes of v96.2 (not officially released) by mro
#
#
# The \textsc{v97.1} release has significant improvements in:
#
# image-generation
# rrm is much faster, requires less memory
# and inline images are aligned more accurately;
#
# image quality --- thanks to Uli Wortmann
# rrm is greatly improved by the use of anti-aliasing effects
# for on-screen clarity, in particular with mathematics,
# text and line-drawings;
#
# memory-requirements
# rrm are much reduced, particularly with image-generation;
#
# mathematics
# rrm can now be handled using a separate parsing procedure;
# images of sub-parts of expressions can be created,
# rather than using a single image for the whole formula;
#
# macro definitions
# rrm having a more complicated structure than previously allowed,
# can now be successfully expanded;
#
# counters and numbering
# rrm are no longer entirely dependent on the \texttt{.aux}
# file generated by \LaTeX;
#
# decisions about which environments
# mro to include or exclude can now be made; see v96.2 .
#
# HTML effects
# rrm for which there is no direct \LaTeX{} counterpart
# can be requested in a variety of new ways;
#
# HTML code
# rrm produced by the translator is much neater and more easily
# readable, containing more comments and fewer redundant breaks
# and \HTMLtag{P} tags.
#
# error-detection
# rrm of simple \LaTeX{} errors, such as missing or unmatched braces,
# is now performed --- a warning message shows a line or two
# of the source code where the error has apparently occurred;
#
#
#
#--------- v96.2 (compiled by jcl)
# mro - Rewrote texexpand. Contains more logic and mimics the
# behaviour of LaTeX more closely. Supports DO_INCLUDE to
# force style file translation and filename extensions in
# (DO|DONT)_INCLUDE.
# mro - Rewrote pstogif and renamed it to pstoimg, as it now also
# produces PNG graphics. Added an automatic configuration
# script (configure-pstoimg). Supplied a patch for dvips-5.58f
# to enable production of EPSF format in multiple files.
# mro - Added configuration variable and command line switch
# LOCAL_ICONS: when set, the navigation icons are copied to
# the document directory.
#
#--------- v96.1 Rev h (compiled by jcl)
# 30-Sep-96
# tak - Fixed bug in optional argument of do_cmd_htmladdimg() in html.perl
# and embed_image(): ALT tag now properly gets transmitted to html.
# hws - uses %new_command and %new_environment instead of /$meta_cmd_rx/
# in process_ext_file()
# - Recover figure and table caption numbers when captions contain
# a ~ [in extract_captions()]
# - environments are not further processed when coming from .aux
# file in process_undefined_environments()
# - introduced translation for \ensuremath to force math expressions
# to be translated into HTML instead of computing an image.
# rrm - introduced latex-only command \etalchar...
# - cite labels fetched from .aux file if available, in do_cmd_bibitem()
# - added call to translate_environments() in do_env_enumerate(),
# do_env_itemize(), do_env_description(), do_env_list().
# - small fix in accent_safe_for_ij()
# - Non-inlined images now embedded in , not
tags.
# - introduced tex2html_nowrap environment.
# LaTeX commands wrapped with this environment go directly into images.tex.
# - introduced make_hbox(), make_vbox() for inlined images
# - check if $adjust is positive in top_justify()
# - changed text_cleanup() to prevent -- => - being applied twice,
# with headings, captions, etc.; thus --- correctly becomes -- not -
# - list items not strongified if $compact, in do_env_description()
# - introduced make_nowrapper(), make_inline_wrapper()
# - \bigskip now produces
# - arguments of \message and \typeout now prompted during translation
# - slight change to process_cmds_in_tex()
# - introduced process_commands_nowrap_in_tex() and
# process_commands_inline_in_tex() to work with tex2html_nowrap resp.
# tex2html_wrap_inline environments.
# Changed list of ignored commands, built a list of nowrap commands.
# - raw arg cmds now also recognized if delimited by @ (eg. a\l@b).
# rrm-nac
# allow new commands \W for non-alphanumerics, not already defined.
# rrm-col
# - introduced $NESTING_LEVEL for counting group ({,} currently) nesting
# - hook for colors used at each section start, in translate()
# jcl - Removed bug in replace_cite_references(), a not executed
# replacement command concerning make_named_href(...).
# - The use of $global{'warnings'}, held in a DBM entry, led to
# subsequent DBM errors when the warnings accumulated to more
# than approx. 900 chars.
# The DBM data base is not used for 'warnings' any more:
# The warnings are now written to the file 'WARNINGS' in the
# directory where all output files reside. The file contents
# is displayed and the file removed at the end of the
# translation process.
# - A \clearpage command precedes the \end{document} of images.tex
# to force the last page to appear in the DVI file. The case
# in which this behaviour is needed is when a style file discards
# the \end{document} command or redefines it (eg. cweb.sty).
# - replaced $(PNMTILE) with $(PBMMAKE) in latex2html.config
# and install-test, changed right_justify() according to the
# suggestion by meh. Removed obsolete blank.pbm.
# - $(GIFTRANS) is preferred to the netpbm binaries if it's available.
# This encreases speed of gif translation significantly if one
# wants to use the (faster) netpbm instead of pbmplus.
# - Changed deal_with_texinputs() to set TEXINPUTS the following way:
# 1. . to let TeX find things from work dir (invocation of l2h)
# or dest dir (for images.tex)
# 2. .. to find things from work dir for images.tex (questionable)
# 3. absolute path to source file
# 4. dest dir, and, 5. $TEXINPUTS (suggested by rrm).
# - Re-arranged latex2html.config to make it more comprehensive:
# - shuffled entries without changing their meaning
# - slight improvements to documentation
# - introduced color variables
# - introduced down url/down title/contents url/index url defaults
# - Tidied l2h's synopsis both in usage() and the nroff section.
# - Introduced l2h options -auto_prefix, -prev_url, -prev_title
# - Introduced global variable $PREAMBLE. It is set to 1 if the text
# before \begin{document} or \startdocument is processed, and 0
# if the text after it is processed. It is *undef* before
# any text is translated.
# - In html.perl, URL in htmladdimg not reverted to raw TeX.
# Klaus Steinberger
# supposed this.
# This is conform with \htmladdnormallink now.
# - Introduced $LINKPOINT (see also latex2html.config), which is used
# to generate a directory index link if the document is finished.
# - Cosmetic changes: removed spaces, indented expressions etc.
# - grep doesn't understand -e on Solaris, fixed install-test
# jcl-env
# Content of environment (\begin{..}...\end{..}) will now be
# encapsulated in l2h open/close brackets internally.
# This will lead to the same behaviour as in TeX: Changes
# within the environment (esp. to font style) are really local
# and reverted when passing the end of the environment.
# jcl-pag
# An incomplete LaTeX command (eg. \theoremstyle{..}) that is
# passed to LaTeX through images.tex leads to an absent page in
# the resulting DVI file, causing all following pages and its
# images to be a page earlier than expected.
# To prevent this, invisible glue is added to each page entry.
# Then, in turn to prevent a page break between glue and overly
# long contents, the \textheight is set to 50cm. This also
# prevents page breaks within the contents (eg. a supertable).
# However, *very* long pages are cut by dvips itself. Anyway, we
# are rid of the 'images out of sync' problem. I'm aware of the
# log entries of Nikos, but couldn't reproduce the restrictions
# he stated.
# To make bug fixes easier an excerpt of the original text is
# added to its translation text in commented form, if $DEBUG.
#--------- v96.1 Rev g (compiled by hws)
# 12-Jun-96
# jcl - Supplied verbatim.perl style file for displaying an external file
# verbatim.
# - LaTeX called from l2h now reads the "images.tex" from
# the current directory before searching TEXINPUTS.
# - Changed &do_cmd_contentsline to be compatible with LaTeX 2.09
# (and 2e).
# jcl-verb
# Extensive changes in the &pre_process routine to handle things
# \verb strings in comments, and commands like \verbatimfile.
# hws - Fixed problem which ocurred when a section label began with
# a number. Allow double and single quotes to appear within
# section headings in segmented documents. Corrected a bug
# in &save_array_in_file which caused an incorrect labels.pl
# to be written. Added a \bodytext{text} command, which inserts
# "text" into the ... preamble of the current
# HTML document page. Depending on the browser, this may allow
# a specification of background and text colours. Added
# an additional "alt=text" parameter to the \htmlimage command to
# override the default ALT HTML tag for figures and tables.
# For LaTeX2e users, added an optional argument to the
# \htmladdimg command for the ALT tag. Upgraded floatfig.perl
# to floatflt.perl to utilize the floatflt package of
# Mats Dahlgren .
# mwk - Made html2e.sty more compatible with the natbib package
# of Patrick W. Daly. Provided natbib.perl and nharvard.perl.
# rrm - Made GIFs used by htmllist.perl transparent and further changes
# to makeidx.perl. The file prefix now applies to images.tex,
# so that images for separate segments are kept separate.
# Hooks were added for packages to modify various links, and to
# to perform operations at the beginning of a document. Added
# sub make_named_href. Improved citation mechanism for
# segmented documents. Implemented mechanism for recognizing
# package options. The command \usepackage[option]{package}
# is now searched in the following order **after** package.perl
# is loaded:
#
# A. file named: _