Contributions to GNU Recode

This is a place where I keep track of my work on GNU Recode.

Contributions:

• RTF filter [rtf-cvs.c (for the current CVS version of Recode)] for Recode. There is an accompanying RTF Testsuite used for testing the RTF->x conversion.
• Baudot filter [baudot-cvs.c (for the current CVS version of Recode)] and some documentation for it, some 5-bit recodings from the turn of the (last) century.
• PETSCII filter [petscii.c (for Recode 3.5) | petscii-cvs.c (for the current CVS version of Recode)] for CBM machines. There is also some documentation on the assignment of Unicode characters for PETSCII.
• Symbol Fonts, Windings etc. (Not even started.)
• Palm Pilot character set mappings Work in progress.

Usage:

For the C-files above. If the file is named xxx-cvs.c, obtain the latest CVS version of recode from the recode homepage at SourceForge (see below). Drop the file xxx-cvs.c in the src/ directory of the source distribution, add xxx-cvs to the variable C_FALLBACKS in the file Makefile.am in the src/ directory.

Then compile as usual (./configure, make, make install).

For the Recode 3.5 compliant files the process is the same, just that you should obtain an Recode 3.4 tarball from the Free Recode Homepage (see below).

PETSCII details

Recode PETSCII <-> UTF-8 with the commands:

recode PETSCII-EN-LC..u8 foo.txt (converts a PETSCII lowercase file named foo.txt to UTF-8)

recode u8..PETSCII-EN-LC foo.txt (converts a UTF-8 file to lowercase PETSCII)

Converting to latin-1 (unix default) or commands like recode pet..l1 foo.txt does not currently work because of some bug in Recode. You can convert to latin-1 (sort of the same as Windows-1252, ISO 8859-1) by first converting to UTF-8 and then convert the result to latin-1 with recode u8/CR..l1 foo.txt. (The /CR after the u8 charset is needed to convert the linefeeds properly.)

Resources:

• Free Recode homepage
• Free Recode at Sourceforge.net
• A brief history of Character Codes for the newbie, written by Steven J. Searle.
• Annotated history of character codes
• The Lorez SZ40/42 cipher attachment - FISH and TUNNY
• Five-unit codes
• Adobe mappings for the Symbol font, using CUS characters.
• Adobe Glyph List and The Complete Adobe Customized (CUS) subarea of Unicode
• An attempt at mapping the Windown Windings font to existing Unicode characters (Caolán McNamara)
• Unicode Technical Reports
  (which also exemplifies the tendency of standard bodies to develop standards for everything that will never be implemented anywhere)
  • Unicode Technical Report #13: Unicode Newline Guidelines
  • Unicode Technical Report #22: Character Mapping Tables
• Palm Pilot Localized applications (info on charsets)
• Chris Tilo

Written by Linus Walleij, /<-r4d s0pHtw4r3.