Third-party components bundled or required at runtime
======================================================

Skald bundles four faces of the Inter typeface family as embedded
assets — InterVariable plus the static-weight Inter-Bold,
Inter-Italic, and Inter-BoldItalic faces. All four ship under the
same SIL Open Font License 1.1; the full licence text travels with
the repository at:

    skald/assets/InterVariable-OFL.txt

Skald also bundles Twemoji-Mozilla — a COLRv0 colour-emoji TTF
built from Twitter's Twemoji artwork (CC-BY-4.0) using Mozilla's
twemoji-colr tooling (Apache-2.0). It's registered as a default
fallback to Inter during `text_init`, so `skald.text("hi 😀")`
renders colour emoji with zero app-side setup.
Attribution + full notice at:

    skald/assets/Twemoji-Mozilla-CCBY.txt

Apps that ship a Skald binary are redistributing the Twemoji
artwork (the bytes are embedded via Odin's compile-time `#load`
regardless of whether the app calls the helper). The CC-BY-4.0
licence requires an attribution line in your app's About / docs:
"Twemoji by Twitter, Inc. and contributors — CC-BY 4.0".

Skald vendors `runa` — a pure-Odin text engine (parsing, shaping,
layout, rasterization, atlas) — at:

    skald/third_party/runa/

Runa is licensed under the zlib license; its own LICENSE file ships
alongside the source at `skald/third_party/runa/LICENSE`. The
vendored tree also includes a set of Unicode UCD reference files at
`skald/third_party/runa/tools/ucd/` — script/category/property data
that runa `#load`s at compile time to build its internal tables,
plus a few conformance test fixtures (ArabicShaping, BidiBrackets,
BidiCharacterTest, DerivedBidiClass, DerivedCoreProperties,
DerivedGeneralCategory, DerivedNormalizationProps, EastAsianWidth,
emoji-data, GraphemeBreakProperty, GraphemeBreakTest,
IndicPositionalCategory, IndicSyllabicCategory, LineBreak,
LineBreakTest, Scripts, SentenceBreakProperty, SentenceBreakTest,
UnicodeData, WordBreakProperty, WordBreakTest). Each file is
© Unicode, Inc. and ships with its own copyright + "for terms of
use and license, see https://www.unicode.org/terms_of_use.html"
header preserved at the top — that header satisfies the
redistribution obligation under Unicode's licence.

Note: Skald's vendored runa intentionally omits the PyThaiNLP
Thai word-break corpus (`thai_words.txt`, CC-BY-SA) that upstream
runa embeds via `linebreak/thai_dict.odin`. Skald replaces that
file with a stub that no-ops the segmenter so every Skald binary
stays permissively licensed by default — apps shipping commercial
products don't have to add a CC-BY-SA attribution line for a
feature most apps don't need. Thai paragraphs fall back to
UAX #14's default SA-class behaviour (one unbreakable run). Apps
that genuinely need Thai word-break segmentation can replace
`skald/third_party/runa/linebreak/thai_dict.odin` with the
upstream version (plus the corpus file) and add the corresponding
CC-BY-SA attribution to their About / docs.

Skald also depends on a number of third-party libraries; none impose
additional obligations on Skald consumers, but they are gratefully
acknowledged below. See README.md for the full list.