This package wraps the utf8proc
library, providing up-to-date inspection and transformation of Unicode
strings and atoms. It is the foundation for SWI-Prolog's normalisation
forms (NFC, NFD, NFKC, NFKD), case folding, grapheme-cluster iteration
(UAX#29), and per-code-point property lookup (general category, bidi
class, display width, combining class, case mappings, ...). The public
Prolog API is exported from library(unicode); see the predicates
unicode_nfc/2, unicode_property/2, atom_graphemes/2, and friends
for details.
A second library, library(unicode_security), ships the
UTS #39 security helpers
(confusable skeleton, mixed-script resolution, identifier restriction
levels) and UAX #24 script
properties. Its tables are derived from the Unicode Character Database
files under data/ by etc/gen_uts39.pl; the generated
uts39_data.{c,h} is checked in. After a UCD version bump, refresh
both with ninja regen-uts39 in the build directory.
This is a git submodule of the main
swi-prolog/swipl-devel
repository. It builds only when the system has libutf8proc installed
(available in every major package manager); the upstream utf8proc source
is no longer vendored. Issues and pull requests specific to this binding
belong here; Unicode-data bugs should be reported upstream at the
utf8proc tracker.