-- $Id: README,v 1.1 2019/11/24 13:21:51 tom Exp $

• Originally added to xterm in 2000 (patch #141), there were a couple of
• updates from Kuhn until 2005 (patch #202), renaming entrypoints and applying
• data from Unicode.org (e.g., 3.2, 4.0, 4.1.0). The Unicode data is
• transformed into tables in this file by a script "uniset" written by Kuhn.
• While Kuhn implemented the original CJK variant, it was unused by xterm
• until Jungshik Shin used it in 2002 to implement the -cjk_width command-line
• option.
• Kuhn added a check for the vertical forms block (double-width) in 2007;
• other updates were derived from the Unicode.org data (release 5.0).

See 00README for Kuhn's original description of the scripts.

The uniset scripts transform the Unicode data into a table of ranges for the character-values, which (with a few special cases) gives xterm a portable method for obtaining character widths. At runtime, xterm can be told to use these tables in preference to (possibly old) system tables of character widths.

When a new release of the Unicode data is available, I get a copy of it. For instance, the last 2019 release was 12.1.0, here:

ftp://ftp.unicode.org/Public/12.1.0/ucd/UCD.zip

Unzipping that into "this" directory, and running

make clean
make

gives these chunks which can be used to replace corresponding text in xterm's wcwidth.c file:

uniset.out: static const struct interval combining[] = {

uniset_cjk.out: static const struct interval ambiguous[] = {

uniset_dbl.out: static const struct interval doublewidth[] = {

uniset_unk.out: static const struct interval unknowns[] = {

NOTE: there are other data files in this directory which are of historical interest, and are retained here because they are less accessible than before.