Fix ``punctuation_chars`` regeneration for tests

Unicode version 7.0 introduced U+2E42, "DOUBLE LOW-REVERSED-9 QUOTATION
MARK", in the Ps category. This codepoint doesn't have a pair
corresponding to it in the Pe category, breaking the assumptions in 
the ``openers`` and ``closers`` strings. The only usages I have been
able to identify are in Old Hungarian, where the character does not
seem to have been used as a quotation mark, so the concept of an
opening character and corresponding closing character do not easily
fit. As the simplest fix to allow tests to pass following
regeneration of the ``punctuation_chars`` file without manual editing,
we remove the U+2E42 character from the Ps category.

git-svn-id: https://svn.code.sf.net/p/docutils/code/trunk@9253 929543f6-e4f2-0310-98a6-ba3bd3dd1d04

1 files changed, 7 insertions, 0 deletions

diff --git a/docutils/tools/dev/generate_punctuation_chars.py b/docutils/tools/dev/generate_punctuation_chars.py
index 11c0108d0..5a7bf9842 100755
--- a/docutils/tools/dev/generate_punctuation_chars.py
+++ b/docutils/tools/dev/generate_punctuation_chars.py
@@ -213,6 +213,13 @@ def character_category_patterns():
     # 301F  LOW DOUBLE PRIME QUOTATION MARK misses the opening pendant:
     ucharlists['Ps'].insert(ucharlists['Pe'].index('\u301f'), '\u301d')
 
+    # 2E42  DOUBLE LOW-REVERSED-9 QUOTATION MARK has no pair, and the only
+    # usages identified thus far are in old hungarian, where it doesn't seem to
+    # be used as a quoting character. Remove from openers (Ps) for now, for
+    # simplicity.
+    # https://www.unicode.org/L2/L2012/12168r-n4268r-oldhungarian.pdf#page=26
+    ucharlists['Ps'].remove('⹂')
+
     # print(''.join(ucharlists['Ps']).encode('utf-8')
     # print(''.join(ucharlists['Pe']).encode('utf-8')
     # print(''.join(ucharlists['Pi']).encode('utf-8')