fix(config): write config.yaml as UTF-8 to stop emoji/personality corruption (#51676)

atomic_yaml_write (and two sibling config writers) called yaml.dump
without allow_unicode=True. The default personalities shipped in cli.py
contain emoji/kaomoji, so PyYAML escaped astral-plane chars as 8-digit
\\UXXXXXXXX sequences inside multi-line double-quoted strings wrapped
with \\ line-continuations. Stricter/non-PyYAML parsers, editors, and
hand-edits break that structure into unclosed quotes, failing the whole
config parse -> silent fallback to defaults -> custom_providers lost.

Add allow_unicode=True to the canonical writer plus tui_gateway/server.py
and the telegram adapter's atomic config write so config is written as
readable UTF-8 with no escape/fold artifacts.

Fixes #51356
This commit is contained in:
Teknium 2026-06-23 23:28:21 -07:00 committed by GitHub
parent 8e7e104521
commit d539cd9004
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 55 additions and 3 deletions

View file

@ -41,3 +41,36 @@ class TestAtomicYamlWrite:
text = target.read_text(encoding="utf-8")
assert "key: value" in text
assert "# comment" in text
def test_writes_unicode_unescaped_and_round_trips(self, tmp_path):
"""Emoji/kaomoji are written as real UTF-8, not fragile escape sequences.
Regression for GitHub #51356: without allow_unicode=True, PyYAML emitted
astral-plane chars (emoji) as 8-digit `\\UXXXXXXXX` escapes inside
multi-line double-quoted strings wrapped with `\\` continuations, which
stricter/non-PyYAML parsers and hand-edits broke into unclosed quotes,
corrupting the entire config.
"""
target = tmp_path / "config.yaml"
# Mirrors the default personalities + skin cursor shipped in cli.py.
data = {
"personalities": {
"kawaii": "kawaii desu~! (◕‿◕) ★ ♪ ヽ(>∀<☆)",
"catgirl": "nya~! (=^・ω・^=) ฅ^•ﻌ•^ฅ",
"surfer": "Cowabunga! 🤙 totally rad bro",
"hype": "LET'S GOOOO!!! 🔥 LEGENDARY!",
},
"display": {"cursor": ""},
}
atomic_yaml_write(target, data)
text = target.read_text(encoding="utf-8")
# No escape artifacts of any kind — real characters on disk.
assert "\\U" not in text
assert "\\u" not in text
# Real glyphs are present verbatim.
assert "🔥" in text
assert "(=^・ω・^=)" in text
# And it reloads to exactly what was written.
assert yaml.safe_load(text) == data