Minor fixes

Author: taminomara

CHANGELOG.md

@@ -13,6 +13,8 @@
 - Added `lua_ls_max_version` config option to safeguard against incompatible changes
   to documentation export format.
 
+- Fixed a few more minor bugs in highlighting of Lua types.
+
 ## [3.4.0]
 
 - **Potential breaking change:** use `confdir` instead of `srcdir` as base path

sphinx_lua_ls/domain.py

@@ -384,7 +384,7 @@ def run(self) -> list[nodes.Node]:
     def parse_signature(self, sig: str) -> tuple[str, T]:
         raise NotImplementedError()
 
-    def use_semicolon_path(self) -> bool:
+    def use_colon_path(self) -> bool:
         return False
 
     def handle_signature_prefix(
@@ -403,7 +403,7 @@ def handle_signature_prefix(
         prefix = "" if classname else ".".join(filter(None, [modname, classname]))
 
         descname = name
-        if self.use_semicolon_path():
+        if self.use_colon_path():
             if "[" in descname:
                 descname_components = utils.separate_sig(descname, ".")
             else:
@@ -696,7 +696,7 @@ def handle_signature(
     def needs_arg_list(self) -> bool:
         return True
 
-    def use_semicolon_path(self) -> bool:
+    def use_colon_path(self) -> bool:
         return self.objtype in ("method", "classmethod")
 
     def get_signature_prefix(
@@ -724,9 +724,6 @@ class LuaData(LuaObject[str]):
     def parse_signature(self, sig):
         name, sig = utils.separate_name_prefix(sig)
 
-        if sig.startswith("=") or sig.startswith(":"):
-            sig = sig[1:]
-
         return name, sig.strip()
 
     @utils.handle_signature_errors
@@ -740,7 +737,15 @@ def handle_signature(
         sw = _SigWriter(signode, self.maximum_signature_line_length)
 
         if typ:
-            sw.punctuation(":")
+            if typ.startswith("="):
+                sw.space()
+                sw.punctuation("=")
+                typ = typ[1:]
+            elif typ.startswith(":"):
+                sw.punctuation(":")
+                typ = typ[1:]
+            else:
+                sw.punctuation(":")
             sw.space()
             sw.typ(typ, self.state.inliner)
 

sphinx_lua_ls/utils.py

@@ -11,7 +11,7 @@
 
 
 #: Regexp for parsing a single Lua identifier.
-_OBJECT_NAME_RE = re.compile(r"^\s*(?P<name>[\w-]+)")
+_OBJECT_NAME_RE = re.compile(r"^\s*(?P<name>[\w-]+)\s*")
 
 #: A single function parameter name.
 _PARAM_NAME_RE = re.compile(r"^\s*[\w-]+\s*$")
@@ -29,7 +29,7 @@ def fn(self, sig: str, signode: addnodes.desc_signature):
                 sig,
                 e,
                 type="lua-ls",
-                location=(signode.source, signode.line),
+                location=signode,
             )
             raise
 
@@ -49,7 +49,7 @@ def separate_name_prefix(sig: str) -> tuple[str, str]:
             name_components.append(f"[{normalize_type(name)}]")
         elif match := _OBJECT_NAME_RE.match(sig):
             name_components.append(match.group("name"))
-            sig = sig[match.span()[1] :]
+            sig = sig[match.end() :]
         else:
             if seen_dot_prefix:
                 raise ValueError("incorrect object name")
@@ -230,32 +230,33 @@ def parse_types(sig: str, parsingFunctionParams: bool = False) -> list[tuple[str
     # Ident not followed by an open brace, semicolon, etc.
     # Example: `module.Type`.
     # Doesn't match: `name?: ...`, `name( ...`, etc.
-    (?P<ident>\w[\w-]*(?:\.\w[\w-]*)*)
-    \s*(?P<ident_qm>\??)\s*
-    (?![:(\w.?-])
+    (?P<ident>\w[\w-]*\s*(?:\.\s*\w[\w-]*\s*)*)
+    (?P<ident_qm>(?:\?\s*)*)
+    (?![:(\w.-])
     |
     # Built-in type not followed by an open brace, semicolon, etc.
     # Example: `string`, `string?`.
     # Doesn't match: `string?: ...`, `string( ...`, etc.
     (?P<type>nil|any|boolean|string|number|integer|function|table|thread|userdata|lightuserdata)
-    \b\s*(?P<type_qm>\??)\s*
-    (?![:(\w.?-])
+    \b\s*(?P<type_qm>(?:\?\s*)*)
+    (?![:(\w.-])
     |
     # Name component, only matches when `ident` and `type` didn't match.
     # Example: `string: ...`.
-    (?P<name>\w[\w.-]*)
+    (?P<name>\w[\w-]*\s*(?:\.\s*\w[\w-]*\s*)*)
     |
     # Punctuation that we separate with spaces.
     (?P<punct>[=:,|&])
     |
     # Braces.
-    (?P<brace>[()[\]{}])
+    # Match them separately, otherwise greedy `other_punct` might eat functional arrow.
+    (?P<brace>[()[\]{}<>])
     |
     # Functional arrow (will be replaced with semicolon).
     (?P<arrow>->)
     |
     # Punctuation that we copy as-is, without adding spaces.
-    (?P<other_punct>[-!#$%*+/\\;<>?@[\]^_{}~]+)
+    (?P<other_punct>[-!#$%*+/\\;?@^_~]+)
     |
     # Anything else is copied as-is.
     (?P<other>.)
@@ -285,6 +286,7 @@ def type_to_nodes(typ: str, inliner) -> list[nodes.Node]:
         elif text := match.group("type"):
             res.append(addnodes.desc_sig_keyword_type(text, text))
             if qm := match.group("type_qm"):
+                qm = re.sub(r"\s", "", qm)
                 res.append(addnodes.desc_sig_punctuation(qm, qm))
         elif text := match.group("string"):
             res.append(addnodes.desc_sig_literal_string(text, text))
@@ -293,14 +295,17 @@ def type_to_nodes(typ: str, inliner) -> list[nodes.Node]:
         elif text := match.group("ident"):
             import sphinx_lua_ls.domain
 
+            text = re.sub(r"\s", "", text)
             ref_nodes, warn_nodes = sphinx_lua_ls.domain.LuaXRefRole()(
                 "lua:_auto", text, text, 0, inliner
             )
             res.extend(ref_nodes)
             res.extend(warn_nodes)
             if qm := match.group("ident_qm"):
+                qm = re.sub(r"\s", "", qm)
                 res.append(addnodes.desc_sig_punctuation(qm, qm))
         elif text := match.group("name"):
+            text = re.sub(r"\s", "", text)
             res.append(addnodes.desc_sig_name(text, text))
         elif text := match.group("punct"):
             if text in "=|&":
@@ -315,7 +320,10 @@ def type_to_nodes(typ: str, inliner) -> list[nodes.Node]:
         elif text := match.group("other_punct"):
             res.append(addnodes.desc_sig_punctuation(text, text))
         elif text := match.group("other"):
-            res.append(nodes.Text(text))
+            if res and isinstance(res[-1], nodes.Text):
+                res[-1] += text
+            else:
+                res.append(nodes.Text(text))
 
     return res
 
@@ -339,17 +347,17 @@ def normalize_type(typ: str) -> str:
         elif text := match.group("type"):
             res += text
             if qm := match.group("type_qm"):
-                res += qm
+                res += re.sub(r"\s", "", qm)
         elif text := match.group("string"):
             res += text
         elif text := match.group("number"):
             res += text
         elif text := match.group("ident"):
-            res += text
+            res += re.sub(r"\s", "", text)
             if qm := match.group("ident_qm"):
-                res += qm
+                res += re.sub(r"\s", "", qm)
         elif text := match.group("name"):
-            res += text
+            res += re.sub(r"\s", "", text)
         elif text := match.group("punct"):
             if text in "=|":
                 res += " "

test/test_regression/doc-directives.html

@@ -2137,13 +2137,6 @@
      data_type_eq
     </span>
    </span>
-   <span class="p">
-    <span class="pre">
-     :
-    </span>
-   </span>
-   <span class="w">
-   </span>
    <span class="w">
    </span>
    <span class="p">