
********** Suggestions of users since Rev g release ***********************


Order:
Contents
Author [Date]


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

I have some problems:

1) The \begin{rawhtml} \end{rawhtml} environment does not seem to work
when specifying funny characters like &#177.
It seems to me that anything like &#<number> is filtered out at some stage.

2) There is extra space added after a  header in the HTML if there was a
\label after the \section in the latex.
This is ugly because it give inconsistent spacing after headers
depending on whether there was a label in the latex.

3) Figures... if a figure has just text in it then why is it first
converted to Postscript, mangled into GIF, and then displayed badly
in a HTML browser?
Cannot figures that contain just formatted text produce the same
formatted text in the HTML output?
I don't see why everything has to be converted to a GIF.

4) The table of contents seems to be ugly to me.
The first entry in the table of contents is....
another table of contents.
I don't see how to have just one contents table.

Tue, 13 Aug 96
Rob Scott <scott@iqsoft.hu>
----------------------------------------------------------------------------


In this version of latex2html:
   $TPATCHLEVEL = "1";
   $TVERSION = "96";
   $RELDATE = "(Feb 5, 1996)";

(I haven't figured out what "revg" means...) install-test prints this:
   Looking for Giftrans...
   Giftrans was found.
   Warning: Cannot find GIFTRANS or GIFTRANS does not accept the -t option.

However, giftrans is installed, and does accept the -t option:
   $giftrans 2>&1 |head -1
   Usage: giftrans [-t color|-T] [-B color] [-b color] [-g oldcolor=newco...

The message is the result of two bugs in install-test.

In look_for, the statement "$what = $_;" assigns to $what the value
from the "type" command.  Under the perl debugger:

     DB<3> p $what
   giftrans is /usr/local/bin/giftrans

The first two tokens, "giftrans is", aren't needed.  Also, $what has
"local" (dynamic) scope, so its value is never used.  After the
subroutine returns, the value of $GIFTRANS has not changed:

     DB<4> p $GIFTRANS
   giftrans

"giftrans" is not in the current directory, so the condition "-x
$GIFTRANS" returns false, so &get_first_line_of_stderr is never
called, and the message is printed.

A fix is to use "which" rather than "type" (eliminates first two
tokens), and to return the full path name via the first element of the
@_ array:

--- install-test-orig   Mon Aug  5 21:56:37 1996
+++ install-test        Sat Aug 10 19:19:11 1996
@@ -178 +178 @@
-       $_ = `sh -c "type  $what"`;
+       $_ = `sh -c "which  $what"`;
@@ -183 +183 @@
-           $what = $_;}
+           $_[0] = $_;}

Jim Van Zandt <jrv@vanzandt.mv.com>
----------------------------------------------------------------------------


I've just tried to convert a tabel with 5 columns from latex to html
using latex2html 96.1-g and got surprising results:
In latex there were 5 columns, but in html only 3.

After a long bug search I found out that latex2html treats \\ and
\newline the same way in the tabular-environment.

This bug can be easy fixed by replacing following line:

in html2.2.pl:
   in sub do_env_tabular {...} :
           @rows = split(/\\\\|\\newline/);
           replaced by:
           @rows = split(/\\\\/);

This helps to behave latex2html the same way as latex.
The \newline-command doesn't bring you to a new row, but to a new line
in the same column.

Rainer Schuetz <schuetzr@informatik.tu-muenchen.de>
----------------------------------------------------------------------------


Well, I browsed through the mailing list archive and found several
references on how to handle the `~' when making a link within the LaTeX
source file, specifically, by using `\~{}', but I have a slightly
different problem.  I'd like to use the command-line option -up_url as in
the following:

darkstar:~/jpl/tex$ latex2html -show_section_numbers -up_url
http://dubhe.cc.nps.navy.mil/~ksford/proposal.html -up_title "Draft
Thesis Proposal" plan

But the problem is that the `~' gets turned into `\%7E'.  First I thought
the ~ was getting expanded by bash, so I used single quotes ('~') to
protect it from the shell.  That didn't work.  Then I tried \~ for the
same reason, also to no avail, then I noticed the commentary on how to
handle these within the LaTeX source, by using \~{} and that didn't work
either.  I even tried protecting this from the shell by using single
quotes around this too:

'\~{}'

But nothing works.

"Kevin S. Ford" <ksford@nps.navy.mil>
----------------------------------------------------------------------------


> Ales Kosir wrote:
> 
> Some suggestions for improvement:
> 
> The behavior of processing babel package changed in 96.1
> version of LaTeX2HTML from 95.1 version.
> The line in LaTeX input file
>   \usepackage[german]{babel}
> is translated in file images.tex into
>   \usepackage[german]{german}
> thus eliminating the use of babel at the processing of images.
> If found this effect rather unexpected. It this is OK for your
> installation of german.sty, Please consider the
> possibility of using babel in connection with other languages, such as
>   \usepackage[english]{babel}
> that also would be transformed into
>   \usepackage[english]{english}
> 
> I suggest to omit the line
>   # $packages = $1 if (/([a-z]+)]{babel}/);
> from file texexpand in future releases.
> 
> More support for non-Latin1 output encoding would also be very welcome.
> 
> I would also like to suggest you if this software can detect LaTeX errors
> while processing images.tex, since this can help detecting LaTeX2HTML
> translation errors.
> 
> A nice notice can also be very helpful if the LaTeX input file
> is newer than .aux file. A new user is surprised that the changes
> introduced in his .tex file does not take effect immediately
> when processing with latex2html.
> ------------------------------------------------

When 96.1 first came out, there was no way to process package options.
Thus the babel package name had to be changed to the language:
french, german, within texexpand in order for the french.perl,
german.perl to be loaded by latex2html.  A side effect is that
the babel package is not used by latex processing the images.

Now there is a way to process package options, so this side effect
should be eliminated.

kosir@lfdtb.fs.uni-lj.si (Ales Kosir)
herb.swan@perc.Arco.com (Swan)
----------------------------------------------------------------------------


I was wondering if anybody out there has succesfully used latex2html on
document which employs the "version" package. I have given an example below.
The translator does not understand two commands from "version.sty"

        includeversion
        excludeversion

I guess that "version.perl" would be fairly simple to write - text between
two "environment-like" delimiters is either kept or discarded!
Unfortunately I can't follow any of the perl scripts that come with
latex2html!

I also guess that the figures are not included because the version
commands are not understood. Would that be a fair assumption?


\documentclass{article}
\usepackage{epsfig}
\usepackage{version}
\begin{document}

\excludeversion{SLIDES}
\includeversion{GUIDE}  % Only includes stuff twixt \begin{GUIDE} & \end{GUIDE}
\excludeversion{NOTES}

hello

\begin{SLIDES}

slides

\epsfig{figure=/home/spm/adamm/HPC/F90Course/Diagrams/LogicalKindDiagram.eps}

\end{SLIDES}

\begin{NOTES}

notes

\epsfig{figure=/home/spm/adamm/HPC/F90Course/Diagrams/LogicalKindDiagram.eps}

\end{verbatim}

\end{NOTES}

\begin{GUIDE}

guide

\epsfig{figure=/home/spm/adamm/HPC/F90Course/Diagrams/LogicalKindDiagram.eps}

\end{GUIDE}

\end{document}

Adam Marshall
Wed, 7 Aug 1996
----------------------------------------------------------------------------


$ latex2html V.tex
This is LaTeX2HTML Version 96.1 (Feb 5, 1996) by Nikos Drakos,
Computer Based Learning Unit, University of Leeds.

OPENING /home/spm/adamm/tmp/V.tex
Cannot create directory V: File exists
Reusing directory V:

Loading /apps5/pdtools/latex2html/styles/texdefs.perl...
Loading /apps5/pdtools/latex2html/styles/epsfig.perl...

Reading ...
Processing macros ...+
Translating ...0/0...........
Writing image file ...

This is TeX, Version 3.1415 (C version 6.1)
(images.tex
LaTeX2e <1994/12/01> patch level 1

Generating postscript images using dvips ...
This is dvipsk 5.526b Copyright 1986, 1993 Radical Eye Software
dvips: ! DVI file can't be opened.
Cannot find file 5999_image002.ps
.giftrans: cannot open img2.gif: No such file or directory
No GIF file!
Cannot find file 5999_image001.ps
.giftrans: cannot open img1.gif: No such file or directory
No GIF file!
Cannot find file 5999_image003.ps
.giftrans: cannot open img3.gif: No such file or directory
No GIF file!

Doing section links ....
Unknown commands: includeversion excludeversion
Done.

--
  Dr AC Marshall                        | Cheese of the Week:
  CSD, University of Liverpool. L69 3BX |  Y Fenni
  Tel: +44-151-794 3722 (Fax: 3759)     |   Even though it's Welsh: Y Fenni

                   http://www.liv.ac.uk/HPC/HPCpage.html
                   http://www.liv.ac.uk/~adamm/HomePage.html

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



> It shouldn't be too difficult to write
> a extension to install-test (I can do it) that checks for the devices
> available in gs.

This is probably a good idea. The way pstogif works right now, you simply
have to set the PNMDEVICE environment variable in the latex2html.config
file and that should work. Install-test could check for the available
devices and change the latex2html.config file accordingly.

Marek Rouchal <marek@btfmd1.fs.uni-bayreuth.de>
>Ok, one could add a path for temporary files (/tmp) where no quotas
>exist...

Marek Rouchal <marek@btfmd1.fs.uni-bayreuth.de>
----------------------------------------------------------------------------


Has anyone else noticed that forced blanks (i.e., "\ ") are translated
incorrectly by latex2html? They were handled correctly in v95.1, as this
version simply translated all occurences of "\ " to " ". Since v95.3
however, the string "\ " is only translated to " " if it does not appear
at the end of a line in the LaTeX source. Obviously, the "\ " string is
translated to "\" (trailing blank is discarded) at a yet-to-be-identified
state in the translation process. As a result, the routine which replaces
"\ " by " " later on will not find those occurences. In the HTML output,
there will be an annoying backslash at the end of that line.

This bug changed at a later revision of v96.1 to the following behaviour:
"\ " never gets translated to " ", thus the HTML output is cluttered with
backslashes. Either in revision F or in revision G, this was fixed, so now
with revision G, we have the same behaviour as in v95.3.

L.Wunner@science-computing.uni-tuebingen.de (Lukas Wunner)
----------------------------------------------------------------------------


One of my colleagues at CERN has some starred versions of the sectioning
commands, like \section*, \subsection*,... whose argument he would like
to include in the table of contents. Therefore he used \addcontentsline,
but the latter (nor \addtocontents) seem to have any effect. latex2html
seems to be quite happy with these commands, however, since it gives not
the slightest hint that it does not know them or cannot handle them
otherwise. Am I missing something?

Michel Goossens <Michel.Goossens@cern.ch>
----------------------------------------------------------------------------


I just upgrade my l2h from ... old version to
revision f (and then revision g). Good.

Usually, my l2h production is stored in a directory called
something like ~myhome/WWW and each translated document <file>
has its corresponding html files stored in
~myhome/WWW/<file>/. Good. (for me...)

But with the last release, I have the surprise that html files
are now store *in* ~myhome/WWW  and not below !!!!
Imagine what happend when l2h asked me if I want to delete or
reuse files in directory... I said yes and all the subdirectories
containing all my l2h job were deleted !!!!

Georges MARIANO <mariano@terre.inrets.fr>
----------------------------------------------------------------------------


I found that the translator spazzed up

\fbox{{\sl Blah, blah, blah}}

but managed 

\fbox{Blah, blah, blah}

"Dr A.C. Marshall" <adamm@liverpool.ac.uk>
----------------------------------------------------------------------------


Dear Steve, and other latex2htmlers,

At 23:02 25/08/1996 -0400, Steve Lawrence wrote:
>When two sections have the same title the numbering is messed up. Are
>there any fixes?
>
>e.g.
>
>\section{Introduction}
>...
>\section{Topic A}
>...
>\subsection{Introduction}

I just came up against this one myself, and seem to have found
a fix.  In the subroutine do_cmd_contentsline, change the line

    eval "\$encoded_${hash}_number{\$title} = \$number";

to

    eval "\$encoded_${hash}_number{\$title} .= \$number . \"$;\"";

From: Sara Hopkins <sh@threel.co.uk> 
      Date: Tue, 27 Aug 1996 17:19:12 +0100 
----------------------------------------------------------------------------


Problem: longtable and supertabular not supported

Solution: added the following to html2.2.pl

# SRL

sub do_env_longtable {
    local($_) = @_;
    &do_env_tabular ($_);
    }

# SRL 

sub do_env_supertabular {
    local($_) = @_;
    &do_env_tabular ($_);
    }

Addendum: is this correct? I think I actually fixed this with preprocessing
w/o testing the above.

[...]
# verbatimtab not supported
# supertabular and longtable not supported
# twocolumn and nips styles cause trouble
# \cal causes problems with image creation
# does not handle \bullet
# equations not centered
# does not understand \$
# multicols environment doesn't work
# does not understand \\*
[...]

1. Tables with two or more tabulars - the caption gets put below the first
tabular but above the rest. e.g. see table 1 in
http://www.neci.nj.nec.com/homepages/lawrence/papers/speech-nips95/l2h/index.html

2. Width of captions equals the width of the table. For very small tables I end up
with:

                                     a    b
                                     10   20
                                     23   50

                                     Caption:
                                     This lo-
                                     oks sill-
                                     y.
                                        
3. When there are two definitions in a document they are both called definition 1.

4. \newcommand{\commandname}{def} is ok but
\newcommand\commandname{def} causes problems.

5. After changing to sffamily, returning to rmfamily is not working always, e.g.
for this latex:

---------

\sffamily

\footnotesize
\begin{samepage}
\begin{verbatimtab}
For each sample vector
        ... removed
\end{verbatimtab} 
\end{samepage}
\normalsize

\rmfamily 

\rmfamily Hence, an image with larger spheres represents increased
violation of the topological ordering of the map. The resulting images
can be used to help with selection of training parameters.

---------

The last sentence is rendered in \sffamily (which I chose to be tt)

Steve Lawrence <*> http://www.neci.nj.nec.com/homepages/lawrence


latex2html: Various problems and some solutions

      From: lawrence@research.nj.nec.com (Steve Lawrence) 
      Date: Thu, 29 Aug 1996 23:56:30 -0400 (EDT) 

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


I have the following LaTeX code:


\subsection{The Generator of Modules \GenoM{}}

\GenoM{} is a generic tool which allows a programmer to specify and
write functional modules in a real-time contro


The \GenoM{} command is treated by TeX and the resulting image is inlined.

In fact it is inlined in the text but not in the section title.

(This used to work in the 91-1 version).

The version I use is:

$TPATCHLEVEL = "1-g";
$TVERSION = "96";
$RELDATE = "(June 11, 1996)";

From: Francois Felix Ingrand <felix@laas.fr> 
      Date: Thu, 5 Sep 1996 10:13:03 +0200 (MET DST) 
----------------------------------------------------------------------------


Hello,

The fact that search engines results or my bookmarks are often
filled with lots of 'Introduction', 'General', ... titles annoyed me.
So I hacked latex2html to support a pre/postfix for the title
tag in all generated html documents. All <H*> tags are unaffected.
(patch against rev-g appended).

        option          config var      default

        -title_prefix   $TITLE_PREFIX   ""
        -title_postfix  $TITLE_POSTFIX  ""

E.g. The \section{Introduction} of the teTeX Guide converted with

        latex2html -title_prefix "teTeX Guide: " TETEXDOC

results in a html title

        <TITLE>teTeX Guide: Introduction</TITLE>

and in an unchanges <H1> tag:

        <H1><A NAME="SECTION00020000000000000000">Introduction</A></H1>

The remain problem is (IMHO) that $TITLE or \title{...} text is also
pre/postfixed.  Anybody know how to fix this?

From: Achim Bohnet <ach@rosat.mpe-garching.mpg.de> 
      Date: Thu, 05 Sep 1996 13:10:29 +0200 
----------------------------------------------------------------------------


Is there a documentation other than the source code on how to extend
the translator's capabilities? If so, where, how, when, ... :o)
I am trying to write the perl-part of the internal style file for
our company and it seems like latex2html ignores the complete text
after the first command that is declared in our local style file.

From: Axel Eble <Axel_Eble@GeNUA.DE> 
      Date: Thu, 05 Sep 1996 17:47:09 +0200 
----------------------------------------------------------------------------


I'm trying to convert a .tex document using latex2html 96.1 and
I'm having a (serious) problem.

After spending some time (1/2 day) on searching for a solution 
in the Manual (troubleshooting), in the mailinglist archive and
the bug reports, this is my last chance for help.

Latex2html stops at "Doing table of contents part" with the
following:

        Doing section links .....
        Doing table of contents ........read error at 
            /usr/local/bin/latex2html line 2564, <INPUT> line 75.

[I've added a new line to be able to place this on this page.]

At line number 2564 in my late2html is the definition of "replace_verb",
i.e. 2562 sub replace_verb {
     2563 # Modifies $_
     2564 s/$verb_mark(\d+)$verb_mark/<code>$verb{$1}<\/code>/go;
     2565 }

From: goce@win.tue.nl (Goce Naumoski) 
      Date: Tue, 17 Sep 1996 14:16:17 +0200 (MET DST) 
----------------------------------------------------------------------------


Is there any support for comment.sty in latex2html?  I can't find it.

Using latex2html 96.1 revg with latex 2.09

We have shared source for design and reference documentation with
conditional inclusion using comment.sty:

\def\moduledesign{\includecomment{design}\excludecomment{reference}}
\def\modulereference{\includecomment{reference}\excludecomment{design}}

\begin{design}
Design text.
\end{design}

(The text in the environment may span several (sub)sections.

Running latex2html produces some warnings:

Translating ...0/57......................1/57......2/57....Cannot find
\end{design}

etc.

From: David Allen <davida@lsl.co.uk> 
      Date: 18 Sep 1996 17:25:31 +0100 
----------------------------------------------------------------------------


I would like latex2html to understand a special comment in the TeXfile like
%LaTeX2HTML: 
in order to automagically determine the options. If you have a different
set of options and a large set of files it would make life easier to include
the options in the TeX-File in a comment, simply run 'latex2html' and the
beast extracts its options all by itself.

From: Axel Eble <Axel_Eble@GeNUA.DE> 
      Date: Thu, 19 Sep 1996 12:47:08 +0200 

NOTE: This can only success partially, as initialisation is done before
      translation! (jcl)

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


 I'm a new LaTeX2HTML user, and I'm very happy such a program exists.
Unfortunately, I've found a number of problems with it. Since I don't
know the first thing about perl, I can't track them down and/or fix
them. I hope others are willing to comment on them.

  1) The alltt environment, when interpreted by l2h, strips off dashes
of long strings of dashes. Since I try to create simple text mode
boxes with dashes (so Lynx users can read them), this is very
inconvenient. For now, I changed the dashes to '=', which does work,
but is uglier.

 2) Also in the alltt environment, I can't display strings like
"<none>". These are passed on as-is to the HTML document, and hence get
interpreted as an unknown command (I presume) and disappear from the
output.

 3) Yet again related to the alltt environment: If a new environment is
created that includes \begin{alltt} ... \end{alltt}, and this
environment is used inside a figure environment, l2h creates an extra
\end{alltt} in the code passed to images.tex, creating serious problems.

 4) Environments that are defined in the document that must be converted
are passed to images.tex expanded. This has a number of consequenses:
 - An environment that uses \newcommand can only be used once (one
   reason to use \newcommand is to transport an environment argument to
   the closing half of the environment). If used more than once, LaTeX
   will complain, when processing images.tex, about the redefinition of
   the command, and ignore it.
 - An environment that uses font-changing commands in the closing half,
   will see them bleed through to the text _after_ the environment.
What reason could there be for passing on the environment in expanded
form? If it would be passed unexpanded, the two above problems wouldn't
occur.

From: Gertjan Klein <gklein@xs4all.nl> 
      Date: Thu, 19 Sep 1996 15:07:50 +0200 (MET DST) 
----------------------------------------------------------------------------


Dear latex2html-ers

A simple list like:

\newcounter{dpart}
\begin{list}{\alph{dpart})}
{\usecounter{dpart}}
 \item this is number a
 \item this is number b
 \item this is number c
\end{list}

Just doesn't work. The letters and ) are left out in the 
html-file. Why?

From: Justin Vervoorn <vervoorn@mkd8.df.op.dlr.de> 
      Date: Fri, 20 Sep 96 11:11:26 METDST 
----------------------------------------------------------------------------


Hi there,

I had problems with pnmcrop: I had some images with a size of about 9 M.
pnmcrop always died because it couldn't allocate enough memory (running
BSDI 2.1 over here with enough RAM, thanks anyway ;o). So I tried mogrify
which is part of ImageMagick and it worked. I then patched pstogif to
use mogrify if it is available and added the appropriate part in
latex2html.config:

$ENV{'MOGRIFY'} = "/usr/X11/bin/mogrify";

This was then added to pstogif:

# Available with ImageMagick
$MOGRIFY=$ENV{'MOGRIFY'} || unset $MOGRIFY;

and later on:

sub crop_scale_etc {
    local($in, $out) = @_;
    local($tmp) = $in . ".tmp";
    open(STDERR, ">/dev/null") unless $DEBUG;

    if ($flip) {
        rename($tmp, $in) unless system("$PNMFLIP -$flip $in > $tmp");
        }
    print("\$in: $in \$tmp: $tmp\n\n");
    # use mogrify if available! 
    if ($MOGRIFY) {
        system("$MOGRIFY -crop 0x0 $in");
        rename($in, $tmp);
    } else {
        system("$PNMCROP $in > $tmp");
    }
...

From: Axel Eble <Axel_Eble@genua.de> 
      Date: Fri, 20 Sep 1996 15:19:52 +0200 
----------------------------------------------------------------------------


