s2t: keep earthly-branch 丑 in 子丑寅卯 (not 醜)

[stantheman0128]

Closes #591.

Problem

STCharacters.txt maps 丑 to 醜 丑, so unmatched 丑 becomes 醜. Earthly-branch phrases such as 丁丑 / 丑时 already have STPhrases.txt self-maps, but 子丑寅卯 (and bare 子丑) did not. Result: 子丑寅卯 -> 子醜寅卯.

Change

Add identity entries in STPhrases.txt:

子丑	子丑
子丑寅卯	子丑寅卯

Do not change the character-level candidate order, so existing ugly-sense conversions (丑陋, 丑闻, 出丑, 他很丑, etc.) stay on 醜 via their phrase / character defaults.

Test plan

• New case BYVoid_OpenCC_Issue_591_chou_earthly_branch in test/testcases/testcases.json
  • expects 子丑寅卯 / 子丑 / 丁丑 / 丑时 to keep earthly-branch 丑
  • expects 丑陋 / 丑闻 / 出丑 / 他很丑 to keep 醜
  • checked under s2t / s2tw / s2twp / s2hk
• Local verification with OpenCC 1.4.1 wheel engine + this branch's text dictionaries (see Evidence)
• CI Bazel dictionary / command_line tests (authoritative)

Evidence

Installed wheel s2t (published dicts, before this fix):

子丑寅卯 -> 子醜寅卯
子丑 -> 子醜
丁丑 -> 丁丑
丑时 -> 丑時
丑陋 -> 醜陋
丑闻 -> 醜聞
出丑 -> 出醜
他很丑 -> 他很醜

This branch, text-dict harness driving data/dictionary/*.txt through the same configs (s2t / s2tw / s2twp / s2hk):

子丑寅卯 -> 子丑寅卯
子丑 -> 子丑
丁丑 -> 丁丑
丑时 -> 丑時
丑陋 -> 醜陋
丑闻 -> 醜聞
出丑 -> 出醜
他很丑 -> 他很醜

sort.py run on STPhrases.txt after the edit (idempotent).

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.

[frankslin]

感謝貢獻。關於這個 ticket，我覺得，最好將所有合法的帶有「丑」的干支組合全部加入。
也要逐個判斷會不會造成特殊歧義。