README.md
In rswipl: Embed 'SWI'-'Prolog'

UCD inputs for the table generators

The files in this directory feed the generators under etc/:

etc/gen_uts39.pl emits the checked-in uts39_data.{c,h} compiled into library(unicode_security).
etc/gen_uniname.pl emits the checked-in compact Unicode name table uniname_data.{c,h} (prototype) from UnicodeData.txt.

The generators run only on demand (ninja regen-uts39 / ninja regen-uniname after a UCD version bump); the .txt files are not checked in.

UnicodeData.txt (for `gen_uniname.pl`)

From https://www.unicode.org/Public/<version>/ucd/:

UnicodeData.txt

UnicodeData.txt carries no version line; pass --version to the generator (the regen-uniname target pins it to match the UCD set).

UTS #39 (Unicode Security Mechanisms) — from https://www.unicode.org/Public/security/<version>/ (or security/latest/ for the most recent release):

confusables.txt
intentional.txt
IdentifierStatus.txt
IdentifierType.txt

UAX #24 (Script_Property, Script_Extensions) — from https://www.unicode.org/Public/<version>/ucd/:

Scripts.txt
ScriptExtensions.txt
PropertyValueAliases.txt (only the sc lines are consumed)

The UTS #39 version may lag the main UCD by one minor release; use the latest matching pair available on unicode.org.

cd packages/utf8proc/data
base_sec=https://www.unicode.org/Public/security/latest
base_ucd=https://www.unicode.org/Public/17.0.0/ucd
for f in confusables.txt intentional.txt \
         IdentifierStatus.txt IdentifierType.txt; do
  curl -sSfO "$base_sec/$f"
done
for f in Scripts.txt ScriptExtensions.txt PropertyValueAliases.txt; do
  curl -sSfO "$base_ucd/$f"
done

After fetching, regenerate from a configured build directory:

ninja regen-uts39

The generator output is deterministic; re-running on unchanged inputs produces a zero diff against the committed uts39_data.{c,h}.

These data files are distributed by Unicode, Inc. under the terms of Unicode Terms of Use and the Unicode data-file license linked from each file's header.