[stantheman0128]

Closes #586.

Problem

Default (non-jieba) MaxMatch matches the shorter STPhrases entry 头发 -> 頭髮 inside 前头发生:

s2t / s2tw / s2twp / s2hk: 前头发生 -> 前頭髮生

Expected (issue body): 前頭發生.

Same A(A/B)(B/C)C overlap pattern as #578 / #581 / #890 / #1304: 前 + 头发 + 生 wins over 前头 + 发生.

Change

Keep 头发 -> 頭髮 (and hair compounds such as 剪头发).

Add one longer STPhrases entry so MaxMatch prefers the reported phrase:

前头发生	前頭發生

Same longer-phrase-wins pattern as #1433 (是只有 vs 是只) and #1440 (人参加 vs 人参). Dictionary sorted with python data/scripts/sort.py data/dictionary/STPhrases.txt.

Test plan

• New cases BYVoid_OpenCC_Issue_586_* in test/testcases/testcases.json
  • reported 前头发生 under s2t / s2tw / s2twp / s2hk
  • hair control: 前头发 / 剪头发 / 头发 still map to 頭髮
  • sibling control: 喉头发干 / 不断发生 unchanged
• Local verification with OpenCC 1.4.x wheel + branch text dictionaries (see Evidence)
• CI Bazel dictionary / command_line tests (authoritative)

Evidence

Repro before (packaged wheel dictionaries = current published behavior):

s2t: 前头发生 -> 前頭髮生
s2t: 前头发 -> 前頭髮
s2t: 头发 -> 頭髮

After this branch (config JSONs with ocd2 refs switched to text, driving data/dictionary/*.txt on the branch):

s2t: 前头发生 -> 前頭發生
s2t: 前头发 -> 前頭髮
s2t: 头发 -> 頭髮
s2t: 剪头发 -> 剪頭髮
s2t: 喉头发干 -> 喉頭發乾
s2t: 不断发生 -> 不斷發生

A/B of existing s2t expectations in testcases.json (branch STPhrases vs expectations, excluding the new #586 cases): 92/92 same, 0 diffs.

sort.py idempotence on STPhrases.txt: no diff.

Please treat CI Bazel tests as authoritative; no local C++/Bazel toolchain was available here.

AI assistance

This fix was developed with AI assistance (Cursor / Grok) and verified locally as shown above.