X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=lib%2Flyx2lyx%2Funicode_symbols.py;h=0c5032acebeb80d0e981cf9cbc3da119a5fe8a61;hb=c24233e3f16e73fbab534deaad51e7a4ff11d8e5;hp=9df65e83f7990f0147ca1c091106a9240d768ea1;hpb=aef6b1681125515418922ce05b269d7d9c4aa693;p=features.git

diff --git a/lib/lyx2lyx/unicode_symbols.py b/lib/lyx2lyx/unicode_symbols.py
index 9df65e83f7..0c5032aceb 100644
--- a/lib/lyx2lyx/unicode_symbols.py
+++ b/lib/lyx2lyx/unicode_symbols.py
@@ -18,44 +18,66 @@
 
 " Import unicode_reps from this module for access to the unicode<->LaTeX mapping. "
 
-import sys, os, re
+import sys, os, re, codecs
+
+# Provide support for both python 2 and 3
+PY2 = sys.version_info[0] == 2
+if not PY2:
+    unichr = chr
+# End of code to support for both python 2 and 3
 
 def read_unicodesymbols():
     " Read the unicodesymbols list of unicode characters and corresponding commands."
-    pathname = os.path.abspath(os.path.dirname(sys.argv[0]))
-    fp = open(os.path.join(pathname.strip('lyx2lyx'), 'unicodesymbols'))
-    spec_chars = []
-    # Two backslashes, followed by some non-word character, and then a character
+    pathname = os.path.abspath(os.path.dirname(__file__))
+    filename = os.path.join(pathname.strip('lyx2lyx'), 'unicodesymbols')
+
+    # Read as Unicode strings in both, Python 2 and 3
+    # Specify the encoding for those systems where the default is not UTF-8
+    fp = codecs.open(filename, encoding="utf8")
+
+    # A backslash, followed by some non-word character, and then a character
     # in brackets. The idea is to check for constructs like: \"{u}, which is how
     # they are written in the unicodesymbols file; but they can also be written
     # as: \"u or even \" u.
-    r = re.compile(r'\\\\(\W)\{(\w)\}')
+    # The two backslashes in the string literal are needed to specify a literal
+    # backslash in the regex. Without r prefix, these would be four backslashes.
+    r = re.compile(r'\\(\W)\{(\w)\}')
+
+    spec_chars = []
     for line in fp.readlines():
-        if line[0] != '#' and line.strip() != "":
-            line=line.replace(' "',' ') # remove all quotation marks with spaces before
-            line=line.replace('" ',' ') # remove all quotation marks with spaces after
-            line=line.replace(r'\"','"') # replace \" by " (for characters with diaeresis)
-            try:
-                [ucs4,command,dead] = line.split(None,2)
-                if command[0:1] != "\\":
-                    continue
-                spec_chars.append([command, unichr(eval(ucs4))])
-            except:
+        if not line.strip() or line.startswith('#'):
+            # skip empty lines and comments
+            continue
+        # Note: backslashes in the string literals with r prefix are not escaped,
+        #       so one backslash in the source file equals one backslash in memory.
+        #       Without r prefix backslahses are escaped, so two backslashes in the
+        #       source file equal one backslash in memory.
+        line=line.replace(' "',' ') # remove all quotation marks with spaces before
+        line=line.replace('" ',' ') # remove all quotation marks with spaces after
+        line=line.replace(r'\"','"') # unescape "
+        line=line.replace(r'\\','\\') # unescape \
+        try:
+            [ucs4,command,dead] = line.split(None,2)
+            if command[0:1] != "\\":
                 continue
-            m = r.match(command)
-            if m != None:
-                command = "\\\\"
-                # If the character is a double-quote, then we need to escape it, too,
-                # since it is done that way in the LyX file.
-                if m.group(1) == "\"":
-                    command += "\\"
-                commandbl = command
-                command += m.group(1) + m.group(2)
-                commandbl += m.group(1) + ' ' + m.group(2)
-                spec_chars.append([command, unichr(eval(ucs4))])
-                spec_chars.append([commandbl, unichr(eval(ucs4))])
+            literal_char = unichr(int(ucs4, 16))
+            if (line.find("notermination=text") < 0 and
+                line.find("notermination=both") < 0 and command[-1] != "}"):
+                command = command + "{}"
+            spec_chars.append([command, literal_char])
+        except:
+            continue
+        m = r.match(command)
+        if m != None:
+            command = "\\"
+            commandbl = command
+            command += m.group(1) + m.group(2)
+            commandbl += m.group(1) + ' ' + m.group(2)
+            spec_chars.append([command, literal_char])
+            spec_chars.append([commandbl, literal_char])
     fp.close()
     return spec_chars
 
 
 unicode_reps = read_unicodesymbols()
+