Add $Language; guard against missing doc (#1497)

A couple of things were noticed during the release: 

* `$Language` should have been added with `Alphabet[]`
* Guard against crashing when a docstring definition is missing.

Author: rocky

mathics/builtin/atomic/strings.py

@@ -238,14 +238,20 @@ class Alphabet(Builtin):
     Some languages are aliases. "Russian" is the same letter set as "Cyrillic"
     >> Alphabet["Russian"] == Alphabet["Cyrillic"]
      = True
+
+    See also <url>
+    :$Language:
+      /doc/reference-of-built-in-symbols/global-system-information/\\$language/
+      </url>.
+
     """
 
     messages = {
         "nalph": "The alphabet `` is not known or not available.",
     }
 
     rules = {
-        "Alphabet[]": """Alphabet["English"]""",
+        "Alphabet[]": """Alphabet[$Language]""",
     }
 
     summary_text = "lowercase letters in an alphabet"

mathics/builtin/system.py

@@ -215,6 +215,58 @@ def eval(self, var, evaluation: Evaluation):
             evaluation.message("GetEnvironment", "name", var)
 
 
+# The current value of $Language
+LANGUAGE = "English"
+
+
+class Language(Predefined):
+    """
+    <url>
+    :WMA link:
+    https://reference.wolfram.com/language/ref/\\$Language.html</url>
+
+    <dl>
+      <dt>'\\$Language'
+      <dd>is a settable global variable for the default language used in Mathics3.
+    </dl>
+
+    See the language in effect used for functions like 'Alphabet[]':
+
+    By setting its value, The letters of 'Alphabet[]' are changed:
+
+    >> $Language = "German"; Alphabet[]
+     = ...
+
+    #> $Language = "English"
+     = English
+
+    See also <url>
+    :Alphabet:
+     /doc/reference-of-built-in-symbols/atomic-elements-of-expressions/string-manipulation/alphabet/
+      </url>.
+    """
+
+    name = "$Language"
+    messages = {
+        "notstr": "`1` is not a string. Only strings can be set as the value of $Language.",
+    }
+
+    summary_text = "settable global variable giving the default language"
+    value = f'"{LANGUAGE}"'
+    # Rules has to come after "value"
+    rules = {
+        "$Language": value,
+    }
+
+    def eval_set(self, value, evaluation: Evaluation):
+        """Set[$Language, value_]"""
+        if isinstance(value, String):
+            evaluation.definitions.set_ownvalue("$Language", value)
+        else:
+            evaluation.message("$Language", "notstr", value)
+        return value
+
+
 class Machine(Predefined):
     """
     <url>:WMA link:https://reference.wolfram.com/language/ref/\\$Machine.html</url>

mathics/doc/doc_entries.py

@@ -643,7 +643,11 @@ def text(self) -> str:
         # used for introspection
         # TODO parse XML and pretty print
         # HACK
+
+        # The following is probably an indication of improper docstring
+        # tagging. But don't crash here.
         if len(self.items) == 0:
+            print(f"Bad doc formatting for {self.title}")
             return "No documentation"
         item = str(self.items[0])
         item = "\n".join(line.strip() for line in item.split("\n"))
@@ -653,7 +657,7 @@ def text(self) -> str:
         item = item.replace("</dt>", "")
         item = item.replace("<dd>", "    ")
         item = item.replace("</dd>", "")
-        item = item.replace("\\$", "_DOLARSIGN_")
+        item = item.replace("\\$", "_DOLLARSIGN_")
         item = (
             item.replace("\\'", "_SINGLEQUOTE_")
             .replace("'", "")
@@ -664,7 +668,7 @@ def text(self) -> str:
 
         item = re.sub(r"\$([0-9a-zA-Z]*)\_\{([0-9a-zA-Z]*)\}\$", r"\1\2", item)
         item = re.sub(r"\$([0-9a-zA-Z]*)\_([0-9a-zA-Z]*)\$", r"\1\2", item)
-        item = item.replace("_DOLARSIGN_", "$")
+        item = item.replace("_DOLLARSIGN_", "$")
         return item