From cd1cd560fdc8ac6396068c2654be39fb0f0ac080 Mon Sep 17 00:00:00 2001
From: Suguru Inatomi <suguru.inatomi@gmail.com>
Date: Wed, 3 Jun 2026 15:51:44 +0900
Subject: [PATCH] fix(lint): allowlist lines with no Japanese characters
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

prh (and other textlint rules) was incorrectly firing on lines that
contain no Japanese characters at all — most notably on English-only
paragraphs left untranslated by the d-additive defer flow during
`/update-origin`. `pnpm lint --fix` would then partially replace
tokens like `Signal Forms` with `シグナルフォーム` inside an
otherwise pure-English sentence, producing nonsensical bilingual
fragments like `シグナルフォームbuilds the field tree...`.

Add a textlint filter allowlist pattern that matches any single line
containing no Hiragana / Katakana / Han characters and excludes it
from rule evaluation. Multi-line English paragraphs are handled
line-by-line, so each pure-English line is skipped while truly
mixed lines (Japanese prose containing `Signal Forms`) still
trigger prh correctly.
---
 .textlintrc | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/.textlintrc b/.textlintrc
index df6ba601cd..602033c3b3 100644
--- a/.textlintrc
+++ b/.textlintrc
@@ -4,7 +4,8 @@
     "allowlist": {
       "allow": [
         "/<docs-code[\\s\\S]*?</docs-code>/m",
-        "/\\{#.*\\}/m"
+        "/\\{#.*\\}/m",
+        "/^[^\\p{Script=Hiragana}\\p{Script=Katakana}\\p{Script=Han}\\n]+$/mu"
       ]
     }
   },