feat(treesitter): add more metadata to `language.inspect()` (#32657)

Problem: No way to check the version of a treesitter parser.

Solution: Add version metadata (ABI 15 parsers only) as well as parser state count and supertype information (ABI 15) in `vim.treesitter.language.inspect()`. Also graduate the `abi_version` field, as this is now the official upstream name.

---------

Co-authored-by: Christian Clason <c.clason@uni-graz.at>

7 files changed, 105 insertions, 37 deletions

diff --git a/runtime/doc/news.txt b/runtime/doc/news.txt
index 0d47c7adb9..b9c1e55305 100644
--- a/runtime/doc/news.txt
+++ b/runtime/doc/news.txt
@@ -409,6 +409,8 @@ TREESITTER
 • |:InspectTree| now shows which nodes are missing.
 • Bundled markdown highlight queries use `conceal_lines` metadata to conceal
   code block fence lines vertically.
+• |vim.treesitter.language.inspect()| shows additional information, including
+  parser version for ABI 15 parsers.
 
 TUI
 
diff --git a/runtime/doc/treesitter.txt b/runtime/doc/treesitter.txt
index 8c63f0b5db..5916ec9371 100644
--- a/runtime/doc/treesitter.txt
+++ b/runtime/doc/treesitter.txt
@@ -1192,14 +1192,17 @@ get_lang({filetype})                      *vim.treesitter.language.get_lang()*
 inspect({lang})                            *vim.treesitter.language.inspect()*
     Inspects the provided language.
 
-    Inspecting provides some useful information on the language like node and
-    field names, ABI version, and whether the language came from a WASM
-    module.
+    Inspecting provides some useful information on the language like ABI
+    version, parser state count (a measure of parser complexity), node and
+    field names, and whether the language came from a WASM module.
 
     Node names are returned in a table mapping each node name to a `boolean`
     indicating whether or not the node is named (i.e., not anonymous).
     Anonymous nodes are surrounded with double quotes (`"`).
 
+    For ABI 15 parsers, also show parser metadata (major, minor, patch
+    version) and a table of supertypes with their respective subtypes.
+
     Parameters: ~
       • {lang}  (`string`) Language
 
diff --git a/runtime/lua/vim/treesitter/_meta/misc.lua b/runtime/lua/vim/treesitter/_meta/misc.lua
index c532257f49..99267bb36e 100644
--- a/runtime/lua/vim/treesitter/_meta/misc.lua
+++ b/runtime/lua/vim/treesitter/_meta/misc.lua
@@ -22,10 +22,15 @@ error('Cannot require a meta file')
 ---@field patterns table<integer, (integer|string)[][]>
 ---
 ---@class TSLangInfo
+---@field abi_version integer
+---@field major_version? integer
+---@field minor_version? integer
+---@field patch_version? integer
+---@field state_count integer
 ---@field fields string[]
 ---@field symbols table<string,boolean>
+---@field supertypes table<string,string[]>
 ---@field _wasm boolean
----@field _abi_version integer
 
 --- @param lang string
 --- @return TSLangInfo
diff --git a/runtime/lua/vim/treesitter/health.lua b/runtime/lua/vim/treesitter/health.lua
index 53b64d1dec..0c9e210fc1 100644
--- a/runtime/lua/vim/treesitter/health.lua
+++ b/runtime/lua/vim/treesitter/health.lua
@@ -35,7 +35,7 @@ function M.check()
     else
       local lang = ts.language.inspect(parsername)
       health.ok(
-        string.format('Parser: %-20s ABI: %d, path: %s', parsername, lang._abi_version, parser)
+        string.format('Parser: %-20s ABI: %d, path: %s', parsername, lang.abi_version, parser)
       )
     end
   end
diff --git a/runtime/lua/vim/treesitter/language.lua b/runtime/lua/vim/treesitter/language.lua
index 38d309a102..f70f99421c 100644
--- a/runtime/lua/vim/treesitter/language.lua
+++ b/runtime/lua/vim/treesitter/language.lua
@@ -168,13 +168,17 @@ end
 
 --- Inspects the provided language.
 ---
---- Inspecting provides some useful information on the language like node and field names, ABI
---- version, and whether the language came from a WASM module.
+--- Inspecting provides some useful information on the language like ABI version, parser state count
+--- (a measure of parser complexity), node and field names, and whether the language came from a
+--- WASM module.
 ---
 --- Node names are returned in a table mapping each node name to a `boolean` indicating whether or
 --- not the node is named (i.e., not anonymous). Anonymous nodes are surrounded with double quotes
 --- (`"`).
 ---
+--- For ABI 15 parsers, also show parser metadata (major, minor, patch version) and a table of
+--- supertypes with their respective subtypes.
+---
 ---@param lang string Language
 ---@return TSLangInfo
 function M.inspect(lang)
diff --git a/src/nvim/lua/treesitter.c b/src/nvim/lua/treesitter.c
index 259f2d4739..fa5cf1118d 100644
--- a/src/nvim/lua/treesitter.c
+++ b/src/nvim/lua/treesitter.c
@@ -263,45 +263,92 @@ int tslua_inspect_lang(lua_State *L)
 
   lua_createtable(L, 0, 2);  // [retval]
 
-  uint32_t nsymbols = ts_language_symbol_count(lang);
-  assert(nsymbols < INT_MAX);
-
-  lua_createtable(L, (int)(nsymbols - 1), 1);  // [retval, symbols]
-  for (uint32_t i = 0; i < nsymbols; i++) {
-    TSSymbolType t = ts_language_symbol_type(lang, (TSSymbol)i);
-    if (t == TSSymbolTypeAuxiliary) {
-      // not used by the API
-      continue;
-    }
-    const char *name = ts_language_symbol_name(lang, (TSSymbol)i);
-    bool named = t != TSSymbolTypeAnonymous;
-    lua_pushboolean(L, named);  // [retval, symbols, is_named]
-    if (!named) {
-      char buf[256];
-      snprintf(buf, sizeof(buf), "\"%s\"", name);
-      lua_setfield(L, -2, buf);  // [retval, symbols]
-    } else {
-      lua_setfield(L, -2, name);  // [retval, symbols]
+  {  // Symbols
+    uint32_t nsymbols = ts_language_symbol_count(lang);
+    assert(nsymbols < INT_MAX);
+
+    lua_createtable(L, (int)(nsymbols - 1), 1);  // [retval, symbols]
+    for (uint32_t i = 0; i < nsymbols; i++) {
+      TSSymbolType t = ts_language_symbol_type(lang, (TSSymbol)i);
+      if (t == TSSymbolTypeAuxiliary) {
+        // not used by the API
+        continue;
+      }
+      const char *name = ts_language_symbol_name(lang, (TSSymbol)i);
+      bool named = t != TSSymbolTypeAnonymous;
+      lua_pushboolean(L, named);  // [retval, symbols, is_named]
+      if (!named) {
+        char buf[256];
+        snprintf(buf, sizeof(buf), "\"%s\"", name);
+        lua_setfield(L, -2, buf);  // [retval, symbols]
+      } else {
+        lua_setfield(L, -2, name);  // [retval, symbols]
+      }
     }
+
+    lua_setfield(L, -2, "symbols");  // [retval]
   }
 
-  lua_setfield(L, -2, "symbols");  // [retval]
+  {  // Fields
+    uint32_t nfields = ts_language_field_count(lang);
+    lua_createtable(L, (int)nfields, 1);  // [retval, fields]
+    // Field IDs go from 1 to nfields inclusive (extra index 0 maps to NULL)
+    for (uint32_t i = 1; i <= nfields; i++) {
+      lua_pushstring(L, ts_language_field_name_for_id(lang, (TSFieldId)i));
+      lua_rawseti(L, -2, (int)i);  // [retval, fields]
+    }
 
-  uint32_t nfields = ts_language_field_count(lang);
-  lua_createtable(L, (int)nfields, 1);  // [retval, fields]
-  // Field IDs go from 1 to nfields inclusive (extra index 0 maps to NULL)
-  for (uint32_t i = 1; i <= nfields; i++) {
-    lua_pushstring(L, ts_language_field_name_for_id(lang, (TSFieldId)i));
-    lua_rawseti(L, -2, (int)i);  // [retval, fields]
+    lua_setfield(L, -2, "fields");  // [retval]
   }
 
-  lua_setfield(L, -2, "fields");  // [retval]
-
   lua_pushboolean(L, ts_language_is_wasm(lang));
   lua_setfield(L, -2, "_wasm");
 
   lua_pushinteger(L, ts_language_abi_version(lang));  // [retval, version]
-  lua_setfield(L, -2, "_abi_version");
+  lua_setfield(L, -2, "abi_version");
+
+  {  // Metadata
+    const TSLanguageMetadata *meta = ts_language_metadata(lang);
+
+    if (meta != NULL) {
+      lua_createtable(L, 0, 3);
+
+      lua_pushinteger(L, meta->major_version);
+      lua_setfield(L, -2, "major_version");
+      lua_pushinteger(L, meta->minor_version);
+      lua_setfield(L, -2, "minor_version");
+      lua_pushinteger(L, meta->patch_version);
+      lua_setfield(L, -2, "patch_version");
+
+      lua_setfield(L, -2, "metadata");
+    }
+  }
+
+  lua_pushinteger(L, ts_language_state_count(lang));
+  lua_setfield(L, -2, "state_count");
+
+  {  // Supertypes
+    uint32_t nsupertypes;
+    const TSSymbol *supertypes = ts_language_supertypes(lang, &nsupertypes);
+
+    lua_createtable(L, 0, (int)nsupertypes);  // [retval, supertypes]
+    for (uint32_t i = 0; i < nsupertypes; i++) {
+      const TSSymbol supertype = *(supertypes + i);
+
+      uint32_t nsubtypes;
+      const TSSymbol *subtypes = ts_language_subtypes(lang, supertype, &nsubtypes);
+
+      lua_createtable(L, (int)nsubtypes, 0);
+      for (uint32_t j = 1; j <= nsubtypes; j++) {
+        lua_pushstring(L, ts_language_symbol_name(lang, *(subtypes + j)));
+        lua_rawseti(L, -2, (int)j);
+      }
+
+      lua_setfield(L, -2, ts_language_symbol_name(lang, supertype));
+    }
+
+    lua_setfield(L, -2, "supertypes");  // [retval]
+  }
 
   return 1;
 }
diff --git a/test/functional/treesitter/language_spec.lua b/test/functional/treesitter/language_spec.lua
index a93b1063a1..e009d97210 100644
--- a/test/functional/treesitter/language_spec.lua
+++ b/test/functional/treesitter/language_spec.lua
@@ -63,7 +63,14 @@ describe('treesitter language API', function()
       return { keys, lang.fields, lang.symbols }
     end))
 
-    eq({ fields = true, symbols = true, _abi_version = true, _wasm = false }, keys)
+    eq({
+      abi_version = true,
+      fields = true,
+      symbols = true,
+      state_count = true,
+      supertypes = true,
+      _wasm = false,
+    }, keys)
 
     local fset = {}
     for _, f in pairs(fields) do