1  Running TeX2page

TeX2page is invoked in much the same way as TeX.1 For instance, given a plain-TeX​ ​document file wherever-it-is/jobname.ext, which can be either a full pathname or a pathname relative to either the working directory or some directory in TEXINPUTS, where jobname is the basename of the file and .ext is its extension, you type

tex  wherever-it-is/jobname. ext

at the operating-system command line.2 You do not need to mention the extension .ext if it is .tex. This creates the output DVI file, jobname.dvi, in the working directory.3

TeX2page is called analogously. To create the HTML version of the same file wherever-it-is/jobname.ext, type

tex2page  wherever-it-is/jobname. ext

Again, the .ext is optional if it is .tex. This creates jobname.html in the working directory.

To try this out, call TeX2page on the example file story.tex provided in all TeX​ ​distributions:

tex2page story

TeX2page will get cracking4 on story.tex, providing the following commentary, or log, on your console:

This is TeX2page, Version 20170129 (ECL 16.1.3 Linux) Sun, Jan 29, 2017, 1:53 am UTC−5
! Missing \end inserted.
Output written on story.html (1 page).

TeX2page is now done, and the result of its labors is the HTML file story.html​ ​(click to see).

The log file story.hlog contains a copy of the above log, and is useful if you didn’t or couldn’t keep track of the console (perhaps because the log was too long). The log says that story.tex lacked a document-ending command such as \end (or \bye) and that TeX2page assumed one anyway. Also, only one HTML page was created, and its name is story.html. TeX2page could in some cases produce auxiliary HTML pages in addition to the main HTML page jobname.html (especially for larger documents). The auxiliary HTML pages are reachable from jobname.html by navigation links (p. 4). As each auxiliary HTML page is completed, the log will show the bracketed numbers [1], [2], etc. The [0] in this log refers to the only HTML file created, viz., story.html.

All this is of course almost exactly analogous to the way you type tex story (or luatex story or xetex story or pdftex story) to get story.dvi (or story.pdf) from story.tex, with the log going into story.log.

This is LuaTeX, Version 1.0.4 (TeX Live 2017/Debian)  (format=luatex 2018.5.10)  3 JUL 2018 07:38
 restricted system commands enabled.
(/usr/share/texlive/texmf-dist/tex/plain/knuth-lib/story.tex [1{/var/lib/texmf/f
Output written on story.pdf (1 page, 35962 bytes).

PDF statistics: 20 PDF objects out of 1000 (max. 8388607)
 13 compressed objects within 1 object stream
 0 named destinations out of 1000 (max. 131072)
 1 words of extra memory for PDF output out of 10000 (max. 100000000)

The only real difference is that TeX​ ​will not add the missing \end on its own, but instead waits for the user to supply it explicitly from the console.5 Note that the bracketed numbers now refer to the pages as numbered in the printed document.

Thus, from one TeX​ ​source file, you can get both a printable .dvi and a browsable .html document, using the same calling conventions.

When TeX​ ​encounters a filename f, it searches for it in a standard list of directories, which can be modified by the user via the environment variable TEXINPUTS. The filename f.tex is tried before f itself is tried. In most modern TeXs, the search is performed using the kpathsea library.

By default, TeX2page will look for files using the same kpathsea mechanism as TeX. However, it is possible to supply a different list of search directories via the environment variable TEX2PAGEINPUTS. It may be useful to have files in TEX2PAGEINPUTS shadow files from TEXINPUTS, because the latter are not really HTML-specific, and can thus be unsuitable for HTML-minded parsing by TeX2page.

In TeXs without the kpathsea library, TEX2PAGEINPUTS is the only way to get TeX2page to automatically access files outside the working directory. Note that TEX2PAGEINPUTS should be a simple list of directory names, colon-separated in Unix and semicolon-separated in Windows. It cannot use the enhanced syntax (viz., * and //) that is typically permitted for TEXINPUTS.

Error recovery in TeX2page is also exactly analogous to TeX, but we will postpone that discussion to p. 11.

Non-file arguments

Like most recent versions of TeX, TeX2page also supports the standard self-identification arguments ‑‑help and ‑‑version. These arguments elicit help only if there isn’t an input file (e.g., ‑‑help.tex) that could match them.

TeX2page called without an argument displays a help message and exits. Unlike TeX, TeX2page will not try to conjure up an input document based purely on console chitchat with an increasingly befuddled user.

In all these cases, the help displayed on the console is also saved in the specially named log file texput.hlog.

Calling TeX2page from Lisp

If, for any reason, it is not possible to call tex2page from your operating-system command line, you may load the file tex2page into your Lisp and call the procedure tex2page:tex2page with the name of the TeX​ ​file as argument:

(load "tex2page") ;use appropriate pathname

(tex2page:tex2page filename)

You can call the procedure tex2page:tex2page several times from the same Lisp session, on the same file or on different files.

Note that we’ve used a package-qualified name, as the Common Lisp TeX2page is provided as a Common Lisp package tex2page whose one exported symbol is tex2page. (But see Appendix A if using the Scheme version of TeX2page.)

Specifying a target directory

By default, TeX2page generates the output HTML files and other auxiliary files (p. 15) in the current working directory. You can tell TeX2page to place its output and auxiliary files in a different directory and thus avoid cluttering up your working directory.

The files used for specifying the target directory are: jobname.hdir in the working directory, .tex2page.hdir in the working directory, and .tex2page.hdir in the user’s HOME directory. The first line of the first of these files that exists is taken to be the name of the target directory. If none of the files exists, the current working directory is the target directory.

For example, if story.hdir contains the filename story as its first line, the HTML and aux files are created in a subdirectory story of the current directory.

The filename may contain the TeX​ ​control sequence \jobname, which expands to the basename of the TeX​ ​document. To always use an auxiliary subdirectory with the same name as the basename of the TeX​ ​document, have ~/.tex2page.hdir contain the line “\jobname” (without quotes).

1 Hereafter, we will use “TeX” to mean any format of TeX, and “plain TeX” when we specifically mean the “plain” format. We will use “Lisp” to refer to the implementation language of the TeX2page script under discussion, although this can be either Common Lisp or Scheme, with near-identical behavior. For the minor differences, see Appendix A (p. 12).

2 The executable tex expects its input file to be marked up in plain TeX. For a LaTeX​ ​document, the executable to use is latex. Modern TeX​ ​engines such as LuaTeX [34], XǝTeX [50], and pdfTeX​ ​have executables named luatex, xetex, pdftex for the plain format, and lualatex, xelatex, and pdflatex for LaTeX.

3 The modern TeXs LuaTeX, XǝTeX, and pdfTeX​ ​produce PDF rather than DVI. In the rest of this manual, whenever we refer to the output DVI file, the reader using the PDF versions of TeX​ ​should read PDF for DVI.

4 TeX2page, like TeX, will find the file story.tex from your distribution as it is in your default TEXINPUTS. If it’s too confusing operating on a file that you don’t see in your immediate space, copy the file manually to your working directory. You can find its location with kpsewhich story.

5 The file story.tex lacks an \end only to demonstrate some interactive capabilities of TeX, which are not relevant for TeX2page.