Merge pull request #363 from tatuylonen/zerozerozero

Greek wiktionary uses '0', '00' and '000' as parameters

Author: xxyzz

src/wikitextprocessor/core.py

@@ -1212,7 +1212,7 @@ def expand_args(coded: str, argmap: TemplateArgs) -> str:
                             expand_args(args[0], argmap), parent, True
                         ).strip()
                         self.expand_stack.pop()
-                        if k.isdigit():
+                        if k.isdigit() and int(k) > 0:
                             k = int(k)
                         else:
                             k = re.sub(r"\s+", " ", k).strip()
@@ -1428,7 +1428,7 @@ def expander(arg: str) -> str:
                             # https://en.wikipedia.org/wiki/Help:Template
                             # (but not around unnamed parameters)
                             k, arg = m2.groups()
-                            if k.isdigit():
+                            if k.isdigit() and int(k) > 0:
                                 k = int(k)
                             else:
                                 self.expand_stack.append("ARGNAME")

src/wikitextprocessor/lua/_sandbox_phase2.lua

@@ -10,22 +10,37 @@ assert(_new_loadData ~= nil)
 
 local function frame_args_index(new_args, key)
     -- print("frame_args_index", key)
-    local i = tonumber(key)
-    if key ~= "inf" and key ~= "nan" and i ~= nil then
-        key = i
-    end
     local v = new_args._orig[key]
     if v == nil then
-        return nil
+        local i = tonumber(key)
+        if i ~= nil then
+            key = i
+        else
+            return nil
+        end
+        v = new_args._orig[key]
+        if v == nil then
+            return nil
+        end
     end
     if not new_args._preprocessed[key] then
         local frame = new_args._frame
-        v = frame:preprocess(v)
+        if type(v) == "userdata" then
+            -- Python tuple in luaexec.call_lua_sandbox.make_frame()
+            local is_named = v[1]
+            v = frame:preprocess(v[0])
+            -- https://en.wikipedia.org/wiki/Help:Template#Whitespace_handling
+            if is_named then
+                v = v:match "^%s*(.-)%s*$"
+            end
+        else
+            v = frame:preprocess(v)
+        end
         -- Cache preprocessed value so we only preprocess each argument once
         new_args._preprocessed[key] = true
         new_args._orig[key] = v
     end
-    -- print("frame_args_index", key, "->", v)
+    -- print("frame_args_index", key, "->", "'"..v.."'")
     return v
 end
 

src/wikitextprocessor/lua/mw.lua

@@ -83,7 +83,90 @@ function mw.clone(v)
 end
 
 function mw.dumpObject(obj)
-    print("mw.dumpObject", obj)
+    -- https://github.com/wikimedia/mediawiki-extensions-Scribunto/blob/184216759e22635bd25a0844e6f68979ecf7bc2a/includes/Engines/LuaCommon/lualib/mw.lua#L551
+    local doneTable = {}
+    local doneObj = {}
+    local ct = {}
+    local function sorter( a, b )
+        local ta, tb = type( a ), type( b )
+        if ta ~= tb then
+            return ta < tb
+        end
+        if ta == 'string' or ta == 'number' then
+            return a < b
+        end
+        if ta == 'boolean' then
+            return tostring( a ) < tostring( b )
+        end
+        return false -- Incomparable
+    end
+    local function _dumpObject( object, indent, expandTable )
+        local tp = type( object )
+        if tp == 'number' or tp == 'nil' or tp == 'boolean' then
+            return tostring( object )
+        elseif tp == 'string' then
+            return string.format( "%q", object )
+        elseif tp == 'table' then
+            if not doneObj[object] then
+                local s = tostring( object )
+                if string.sub(s, 1, 6) == 'table:' then  -- this line changed
+                    ct[tp] = ( ct[tp] or 0 ) + 1
+                    doneObj[object] = 'table#' .. ct[tp]
+                else
+                    doneObj[object] = s
+                    doneTable[object] = true
+                end
+            end
+            if doneTable[object] or not expandTable then
+                return doneObj[object]
+            end
+            doneTable[object] = true
+
+            local ret = { doneObj[object], ' {\n' }
+            local mt = getmetatable( object )
+            local indentString = "  "
+            if mt then
+                ret[#ret + 1] = string.rep( indentString, indent + 2 )
+                ret[#ret + 1] = 'metatable = '
+                ret[#ret + 1] = _dumpObject( mt, indent + 2, false )
+                ret[#ret + 1] = "\n"
+            end
+
+            local doneKeys = {}
+            for key, value in ipairs( object ) do
+                doneKeys[key] = true
+                ret[#ret + 1] = string.rep( indentString, indent + 2 )
+                ret[#ret + 1] = _dumpObject( value, indent + 2, true )
+                ret[#ret + 1] = ',\n'
+            end
+            local keys = {}
+            for key in pairs( object ) do
+                if not doneKeys[key] then
+                    keys[#keys + 1] = key
+                end
+            end
+            table.sort( keys, sorter )
+            for i = 1, #keys do
+                local key = keys[i]
+                ret[#ret + 1] = string.rep( indentString, indent + 2 )
+                ret[#ret + 1] = '['
+                ret[#ret + 1] = _dumpObject( key, indent + 3, false )
+                ret[#ret + 1] = '] = '
+                ret[#ret + 1] = _dumpObject( object[key], indent + 2, true )
+                ret[#ret + 1] = ",\n"
+            end
+            ret[#ret + 1] = string.rep( indentString, indent )
+            ret[#ret + 1] = '}'
+            return table.concat( ret )
+        else
+            if not doneObj[object] then
+                ct[tp] = ( ct[tp] or 0 ) + 1
+                doneObj[object] = tostring( object ) .. '#' .. ct[tp]
+            end
+            return doneObj[object]
+        end
+    end
+    return _dumpObject( obj, 0, true )
 end
 
 function mw.incrementExpensiveFunctionCount()

src/wikitextprocessor/luaexec.py

@@ -135,7 +135,7 @@ def recurse(x: Union[list, tuple, dict]) -> Any:
         # Convert numeric keys to integers and see if we can make it a
         # table with sequential integer keys.
         for k, v in list(x.items()):
-            if k.isdigit():
+            if k.isdigit() and int(k) > 0:
                 del x[k]
                 x[int(k)] = recurse(v)
             else:
@@ -429,29 +429,31 @@ def make_frame(
             frame_args = {}
             for k, arg in args.items():
                 arg = re.sub(r"(?si)(<\s*noinclude\s*/\s*>|\n$)", "", arg)
-                frame_args[k] = arg
+                frame_args[k] = (arg, False)
         else:
             assert isinstance(args, (list, tuple))
             frame_args = {}
             num = 1
             for arg in args:
                 # |-separated strings in {{templates|arg=value|...}}
                 m = re.match(r"""(?s)^\s*([^<>="']+?)\s*=\s*(.*?)\s*$""", arg)
-                if m:
-                    # Have argument name
+                if m is not None:
+                    # named parameter
                     k, arg = m.groups()
-                    if k.isdigit():
+                    if k.isdigit() and int(k) > 0:
+                        # Greek wiktionary uses '0', '00' and '000' as
+                        # parameter names...
                         k = int(k)
-                        if k < 1 or k > 1000:
+                        if k > 1000:
                             ctx.warning(
-                                f"Template argument index <1 or >1000: {k=!r}",
+                                f"Template argument index >1000: {k=!r}",
                                 sortid="luaexec/477/20230710",
                             )
                             k = 1000
                         if num <= k:
                             num = k + 1
                 else:
-                    # No argument name
+                    # unnamed parameter
                     k = num
                     num += 1
                 if k in frame_args:
@@ -465,7 +467,7 @@ def make_frame(
                 # does not always like them (e.g., remove_links() in
                 # Module:links).
                 arg = re.sub(r"(?si)(<\s*noinclude\s*/\s*>|\n$)", "", arg)
-                frame_args[k] = arg
+                frame_args[k] = (arg, m is not None)
         frame_args_lt: "_LuaTable" = lua.table_from(frame_args)  # type: ignore[union-attr]
 
         def extensionTag(frame: "_LuaTable", *args: Any) -> str:

src/wikitextprocessor/parser.py

@@ -632,7 +632,10 @@ def template_parameters(self) -> TemplateParameters:
                             parameter_value = parameter[
                                 equal_sign_index + 1 :
                             ].strip()
-                            if parameter_name.isdigit():  # value contains "="
+                            if (
+                                parameter_name.isdigit()
+                                and int(parameter_name) > 0
+                            ):  # value contains "="
                                 parameter_name = int(parameter_name)
                                 is_named = False
                             if len(parameter_value) > 0:

tests/test_lua.py

@@ -598,3 +598,45 @@ def test_mw_uri_anchorEncode(self):
 return export""",
         )
         self.assertEqual(self.wtp.expand("{{#invoke:test|test}}"), "**")
+
+    def test_el_zero_arg(self):
+        # https://el.wiktionary.org/wiki/Πρότυπο:ετ
+        # Unnamed template parameters and numbered parameters can only
+        # be positive non-zero integers; zero or "00" or negative is a string
+        self.wtp.start_page("θηλυκός")
+        self.wtp.add_page(
+            "Module:test",
+            828,
+            """local export = {}
+function export.test(frame)
+  return tostring(frame.args['0']) .. "|" ..
+         --tostring(frame.args[0]) .. "|" ..
+         tostring(frame.args['00']) .. "|" ..
+         tostring(frame.args[1]) .. "|" ..
+         tostring(frame.args[2]) .. "|" ..
+         tostring(frame.args['named'])
+end
+return export""",
+        )
+        self.assertEqual(
+            self.wtp.expand(
+                "{{#invoke:test|test|0= 0 |00= 00 | first |2= second |named= named }}"  # noqa: E501
+            ),
+            "0|00| first |second|named",
+        )
+
+    def test_el_strip_arg(self):
+        self.wtp.start_page("θηλυκός")
+        self.wtp.add_page(
+            "Module:test",
+            828,
+            """local export = {}
+function export.test(frame)
+  return tostring(frame.args['foo'])
+end
+return export""",
+        )
+        self.assertEqual(
+            self.wtp.expand("{{#invoke:test|test|foo=  {{#if||}} {{#if||}} }}"),
+            "",
+        )