db2x_texixml — Make Texinfo files from Texi-XML
db2x_texixml
[options...] [xml-document
]
db2x_texixml converts a Texi-XML document into one or more Texinfo documents.
If xml-document
is not given, then the document
to convert comes from standard input.
The filenames of the Texinfo documents are determined by markup in the Texi-XML source. (If the filenames are not specified in the markup, then db2x_texixml attempts to deduce them from the name of the input file. However, the Texi-XML source should specify the filename, because it does not work when there are multiple output files or when the Texi-XML source comes from standard input.)
--encoding=encoding
Select the character encoding used for the output files.
The available encodings are those of
iconv.
The default encoding is us-ascii
.
The XML source may contain characters that are not representable in the encoding that you select; in this case the program will bomb out during processing, and you should choose another encoding. (This is guaranteed not to happen with any Unicode encoding such as UTF-8, but unfortunately not everyone is able to process Unicode texts.)
If you are using GNU’s version of
iconv, you can affix
//TRANSLIT
to the end of the encoding name
to attempt transliterations of any unconvertible characters in the output.
Beware, however, that the really inconvertible characters will be turned
into another of those damned question marks. (Aren’t you sick of this?)
The suffix //TRANSLIT
applied
to a Unicode encoding — in particular, utf-8//TRANSLIT
—
means that the output files are to remain in Unicode,
but markup-level character translations using utf8trans
are still to be done. So in most cases, an English-language
document, converted using
--encoding=
will actually end up as a US-ASCII document,
but any untranslatable characters
will remain as UTF-8 without any warning whatsoever.
(Note: strictly speaking this is not “transliteration”.)
This method of conversion is a compromise over strict
utf-8//TRANSLIT
--encoding=
processing, which aborts if any untranslatable characters are
encountered.
us-ascii
Note that man pages and Texinfo documents
in non-ASCII encodings (including UTF-8)
may not be portable to older (non-internationalized) systems,
which is why the default value for this option is
us-ascii
.
To suppress any automatic character mapping or encoding conversion
whatsoever, pass the option
--encoding=
.
utf-8
--list-files
Write a list of all the output files to standard output, in addition to normal processing.
--output-dir=dir
Specify the directory where the output files are placed. The default is the current working directory.
This option is ignored if the output is to be written
to standard output (triggered by the
option --to-stdout
).
--to-stdout
Write the output to standard output instead of to individual files.
If this option is used even when there are supposed to be multiple output documents, then everything is concatenated to standard output. But beware that most other programs will not accept this concatenated output.
This option is incompatible with --list-files
,
obviously.
--info
Pipe the Texinfo output to makeinfo, creating Info files directly instead of Texinfo files.
--plaintext
Pipe the Texinfo output to makeinfo
--no-headers
, thereby creating
plain text files.
--help
Show brief usage information and exit.
--version
Show version and exit.
This program uses certain other programs for its operation. If they are not in their default installed locations, then use the following options to set their location:
--utf8trans-program=path
, --utf8trans-map=charmap
Use the character map charmap
with the utf8trans program, included with docbook2X, found
under path
.
--iconv-program=path
The location of the iconv program, used for encoding conversions.
Texinfo language compatibility. The Texinfo files generated by db2x_texixml sometimes require Texinfo version 4.7 (the latest version) to work properly. In particular:
db2x_texixml relies on makeinfo to automatically add punctuation after a @ref if it it not already there. Otherwise the hyperlink will not work in the Info reader (although makeinfo will not emit any error).
The new @comma{} command is used for commas
(,
) occurring inside argument lists to
Texinfo commands, to disambiguate it from the comma used
to separate different arguments. The only alternative
otherwise would be to translate ,
to
.
which is obviously undesirable (but earlier docbook2X versions
did this).
If you cannot use version 4.7 of makeinfo, you can still use a sed script to perform manually the procedure just outlined.
Relation of Texi-XML with the XML output format of makeinfo.
The Texi-XML format used by docbook2X is different
and incompatible with the XML format generated by
makeinfo
with its --xml
option.
This situation arose partly because the Texi-XML format
of docbook2X was designed and implemented independently
before the appearance
of makeinfo’s XML format.
Also Texi-XML is very much geared towards being
machine-generated from other XML formats,
while there seems to be no non-trivial applications
of makeinfo’s XML format.
So there is no reason at this point for docbook2X
to adopt makeinfo’s XML format
in lieu of Texi-XML.
Text wrapping in menus is utterly broken for non-ASCII text. It is probably also broken everywhere else in the output, but that would be makeinfo’s fault.
--list-files
might not work correctly
with --info
. Specifically, when the output
Info file get too big, makeinfo will decide
to split it into parts named
,
abc
.info-1
,
abc
.info-2
, etc.
db2x_texixml does not know exactly how many of these files
there are, though you can just do an ls
to find out.
abc
.info-3