Merge a40031b into b79276c

sthibaul · web-flow · commit 36cd65d3f636 · 2020-06-11T20:12:09.000+02:00
diff --git a/devDocs/developerGuide.t2t b/devDocs/developerGuide.t2t
@@ -72,6 +72,7 @@ Subsequent lines contain a textual identifier used to identify the symbol, a tab
 For example:
 ```
 . sentence ending	(?<=[^\s.])\.(?=[\"')\s]|$)
+dates .	\b(\d\d)\.(\d\d)\.(\d{2}|\d{4})\b
 ```
 
 Again, the English symbols are inherited by all other locales, so you need not include any complex symbols already defined for English.
@@ -98,6 +99,8 @@ Certain characters cannot be typed into the file, so the following special seque
  - \f: form feed
  - \#: # character (needed because # at the start of a line denotes a comment)
 - replacement: The text which should be spoken for the symbol.
+If the symbol is a complex symbol, \1, \2, etc. can be used to refer to the groups matches, which will be inlined in the replacement, allowing for simpler rules.
+This also means that to get a \ character in the replacement, one has to type \\.
 - level: The symbol level at which the symbol should be spoken.
 The symbol level is configured by the user and specifies the amount of symbols that should be spoken.
 This field should contain one of the levels "none", "some", "most", "all" or "char", or "-" to use the default.
@@ -133,6 +136,13 @@ It means that the ". sentence ending" complex symbol should be spoken as "point"
 Level and preserve are not specified, so they will be taken from English.
 A display name is provided so that French users will know what the symbol represents.
 
+```
+dates .	\1 \2 \3	all	always	# point de date
+```
+This line appears in the French symbols.dic file.
+It means that the first, second, and third groups of the match will be included, separated by spaces.
+The effect is thus to remove the dots from the date.
+
 Please see the file locale\en\symbols.dic for the English definitions which are inherited for all locales.
 This is also a good full example.
 
diff --git a/source/characterProcessing.py b/source/characterProcessing.py
@@ -540,15 +540,34 @@ def _regexpRepl(self, m):
 			if group == "simple":
 				# Simple symbol.
 				symbol = self.computedSymbols[text]
+				replacement = symbol.replacement
 			else:
 				# Complex symbol.
 				index = int(group[1:])
 				symbol = self._computedComplexSymbolsList[index]
+				replacement = ''
+
+				escape = False
+				for c in symbol.replacement:
+					if not escape:
+						if c == '\\':
+							escape = True
+						else:
+							replacement += c
+					else:
+						if c == '\\':
+							replacement += '\\'
+						elif c >= '0' and c <= '9':
+							replacement += m.group(m.lastindex + ord(c) - ord('0'))
+						else:
+							log.error("Invalid reference \\%s" % c)
+						escape = False
+
 			if symbol.preserve == SYMPRES_ALWAYS or (symbol.preserve == SYMPRES_NOREP and self._level < symbol.level):
 				suffix = text
 			else:
 				suffix = " "
-			if self._level >= symbol.level and symbol.replacement:
+			if self._level >= symbol.level and replacement:
 				return u" {repl}{suffix}".format(repl=symbol.replacement, suffix=suffix)
 			else:
 				return suffix
diff --git a/tests/unit/test_characterProcessing.py b/tests/unit/test_characterProcessing.py
@@ -0,0 +1,21 @@
+#tests/unit/test_characterProcessing.py
+#A part of NonVisual Desktop Access (NVDA)
+#This file is covered by the GNU General Public License.
+#See the file COPYING for more details.
+#Copyright (C) 2020 NV Access Limited
+
+"""Unit tests for the characterProcessing module.
+"""
+
+import unittest
+from characterProcessing import SYMLVL_ALL
+from characterProcessing import processSpeechSymbols as process
+
+class TestBasic(unittest.TestCase):
+
+	# TODO: check that complex symbols work, check that levels work, check
+	# that fallback to english works
+
+	def test_FR(self):
+		"""Test some French replacements"""
+		self.assertEqual(process("fr_FR", "-", SYMLVL_ALL), "tiret")
