If you click on [I] in the web browser or press [I]
in the text mode interface or say
gms -i respectively
gms /i in the command
line, then the TeX engine will initialize a
so-called format file for Markup Shredder, which is given the base
name gerolf and the extension efmt (Linux, Windows 32) or efm (Dos) in case of the pdfetex engine, and
fmt for pdftex. In binary format, this file
contains information taken from the following program texts or data structures,
as reported in gerolf.log:
prologue.cfg: Commands to be executed before
initializing Markup Shredder.plain.texplain.tex: The well-tested plain TeX macros (44 kB)
by Donald Ervin Knuth, as described in appendix B of The
TeXbook (1984).gerolf.tex and other TeX files in
[GMS_ROOT]/tex/gerolf: The Markup Shredder TeX
macros (about 300 kB) by G. D. Brettschneider.typeset.cfg: User-editable typesetting parameters, as
described in The TeXbook,
pages 272–275, 348–349 and 451.alias.cfg: User-defined font names, as described in
the write font map chapter.markup.tex: Definition of supported HTML elementselements (or “tags”),
attributesattributes and CSS propertiesproperties, as listed in
gerolf.log, section 1d.plugin.cfg: List of installed data structures.font.map: List of fonts, as described in the write
font map chapter; see gerolf.log, section 2a.phv.krn and other margin kerning files for character protruding of selected fonts
in [GMS_ROOT]/data/krn; see
gerolf.log, section 2b.en-US.texen-US.tex
(originally named hyphen.tex, by Donald Ervin Knuth) and other
hyphenation pattern lists for various languages in
[GMS_ROOT]/tex/hyphen; see
gerolf.log, section 2c.
These pattern files, downloaded from CTAN and re-encoded
according to code pages supported by GMS, are human-readable. Maybe you can
contribute a new one for your native language – or check the
existing file? Here’s a hint for understanding them: “A hyphen is
considered to be acceptable between two letters if the associated interletter
value is odd. A large odd value forces desirable hyphen points, a large even
value suppresses undesirable hyphens, for example:
.hy3phe2n5at2io2n.”
(The TeXbook,
page 450).
If you discover wrong hyphens in the GMS output PDF, you can define an
exception list at the end of the pattern file for your language, like
this one: \hyphenation {man-u-script
ap-pen-dix}.u0000.row and other Unicode row definitions (see
gerolf.log, section 2d), which are excerpts of 256 code
points from the character space as specified in UnicodeData.txt. This database file is described in
UCD.html.
If you want to use fonts with alphabets that are not supported initially by
GMS, say Tibetan, then you have to extract the required u0F00.row
file by yourself, comprising characters on Unicode points U+0F00
to U+0FFF. GMS then will produce u0F00.enc, the
required encoding file to support a font with Unicode glyph names like
/uni0700 etc. It is up to you to provide such a font.HTMLlat1.entHTMLlat1.ent and other definition files for named
entities or character references in markup documents (like
ß for German ß), as defined in the
HTML 4.01HTML 4.01 specification; see
gerolf.log, section 2e.gms.gly and other glyph lists associating Unicode
points with PostScript glyph names. A lot of Unicode characters were given
multiple glyph names while encoding files must decide for a
single; so gms.gly is a compromise on the Adobe and Windows glyph lists. If encoding files with other glyph names like
arabicmt.enc are needed, they must be created
manually.
In gerolf.log, section 2f, you may read warnings like
!Undefined unicode row 3000, indicating that
the corresponding file u3000.rowu3000.row (Hiragana, Katakana) was moved to
[GMS_ROOT]/data/row/_disable, because GMS does not
provide a matching font.cp1252.txt (West Europe) and definition files for other
code pages which are recognized by Markup Shredder (see
gerolf.log, section 2g), if the <head>
element of a markup file contains a tag like
<meta http-equiv =
"content-type" content =
"text/html; charset=windows-1252" />.
These files, representing a selection of Unicode characters numbered
arbitrarily from 0 to 255, are used by GMS to extract PostScript glyph names
from gms.gly when writing encoding files like cp1252.enc.epilogue.cfg: Commands that shall be executed after
initializing Markup Shredder, for example to generate encoding files.
You have to re-initialize the TeX format file for GMS whenever you made changes to the
information mentioned above (for instance by adding hyphenation patterns for
another language, or by adding new fonts and re-writing the
font.map), or when you have downloaded a newer pdfetex
binary for Windows or Linux, together with the corresponding message pool file, from Sebastian Rahtz’s TeXlive distribution.
There is a mailing list where problems concerning the
pdfetex binary can be discussed. Please do not expect members
of this list to be familiar with Markup Shredder.