refactor(treesitter): use scratch buffer for string parser #35988

This commit changes `languagetree.lua` so that it creates a scratch
buffer under the hood when dealing with string parsers. This will make
it much easier to just use extmarks whenever we need to track injection
trees in `languagetree.lua`. This also allows us to remove the
`treesitter.c` code for parsing a string directly.

Note that the string parser's scratch buffer has `set noeol nofixeol` so
that the parsed source exactly matches the passed in string.

runtime/doc/treesitter.txt

@@ -1468,10 +1468,10 @@ LanguageTree:register_cbs({cbs}, {recursive})
                      callbacks.
 
 LanguageTree:source()                                  *LanguageTree:source()*
-    Returns the source content of the language tree (bufnr or string).
+    Returns the source bufnr of the language tree.
 
     Return: ~
-        (`integer|string`)
+        (`integer`)
 
                                                *LanguageTree:tree_for_range()*
 LanguageTree:tree_for_range({range}, {opts})

runtime/lua/vim/treesitter/_meta/misc.lua

@@ -5,7 +5,7 @@ error('Cannot require a meta file')
 ---@alias TSLoggerCallback fun(logtype: 'parse'|'lex', msg: string)
 
 ---@class TSParser: userdata
----@field parse fun(self: TSParser, tree: TSTree?, source: integer|string, include_bytes: boolean, timeout_ns: integer?): TSTree?, (Range4|Range6)[]
+---@field parse fun(self: TSParser, tree: TSTree?, source: integer, include_bytes: boolean, timeout_ns: integer?): TSTree?, (Range4|Range6)[]
 ---@field reset fun(self: TSParser)
 ---@field included_ranges fun(self: TSParser, include_bytes: boolean?): integer[]
 ---@field set_included_ranges fun(self: TSParser, ranges: (Range6|TSNode)[])

runtime/lua/vim/treesitter/languagetree.lua

@@ -99,7 +99,8 @@ local TSCallbackNames = {
 ---taken from _trees. This is mostly a short-lived cache for included_regions()
 ---@field private _lang string Language name
 ---@field private _parent? vim.treesitter.LanguageTree Parent LanguageTree
----@field private _source (integer|string) Buffer or string to parse
+---@field private _source integer Buffer to parse
+---@field private _has_scratch_buf boolean Whether _source is a |scratch-buffer| for string parsing.
 ---@field private _trees table<integer, TSTree> Reference to parsed tree (one for each language).
 ---Each key is the index of region, which is synced with _regions and _valid.
 ---@field private _valid_regions table<integer,true> Set of valid region IDs.
@@ -134,11 +135,26 @@ function LanguageTree.new(source, lang, opts)
     source = vim.api.nvim_get_current_buf()
   end
 
+  local has_scratch_buf = false
+
+  if type(source) == 'string' then
+    local new_source = vim.api.nvim_create_buf(false, true)
+    if new_source == 0 then
+      error('Unable to create buffer for string parser')
+    end
+    vim.bo[new_source].fixeol = false
+    vim.bo[new_source].eol = false
+    vim.api.nvim_buf_set_lines(new_source, 0, -1, false, vim.split(source, '\n', { plain = true }))
+    source = new_source
+    has_scratch_buf = true
+  end
+
   local injections = opts.injections or {}
 
   --- @class vim.treesitter.LanguageTree
   local self = {
     _source = source,
+    _has_scratch_buf = has_scratch_buf,
     _lang = lang,
     _children = {},
     _trees = {},
@@ -174,8 +190,7 @@ end
 
 --- @private
 function LanguageTree:_set_logger()
-  local source = self:source()
-  source = type(source) == 'string' and 'text' or tostring(source)
+  local source = tostring(self:source())
 
   local lang = self:lang()
 
@@ -365,8 +380,8 @@ function LanguageTree:children()
   return self._children
 end
 
---- Returns the source content of the language tree (bufnr or string).
---- @return integer|string
+--- Returns the source bufnr of the language tree.
+--- @return integer
 function LanguageTree:source()
   return self._source
 end
@@ -515,9 +530,8 @@ function LanguageTree:_async_parse(range, on_parse)
   end
 
   local source = self._source
-  local is_buffer_parser = type(source) == 'number'
-  local buf = is_buffer_parser and vim.b[source] or nil
-  local ct = is_buffer_parser and buf.changedtick or nil
+  local buf = vim.b[source]
+  local ct = buf.changedtick
   local total_parse_time = 0
   local redrawtime = vim.o.redrawtime * 1000000
 
@@ -527,10 +541,7 @@ function LanguageTree:_async_parse(range, on_parse)
   local parse = coroutine.wrap(self._parse)
 
   local function step()
-    if is_buffer_parser then
-      if
-        not vim.api.nvim_buf_is_valid(source --[[@as number]])
-      then
+    if not vim.api.nvim_buf_is_valid(source) then
       return nil
     end
 
@@ -540,7 +551,6 @@ function LanguageTree:_async_parse(range, on_parse)
       total_parse_time = 0
       parse = coroutine.wrap(self._parse)
     end
-    end
 
     thread_state.timeout = not vim.g._ts_force_sync_parsing and default_parse_timeout_ns or nil
     local parse_time, trees, finished = tcall(parse, self, range, thread_state)
@@ -725,6 +735,10 @@ end
 --- `remove_child` must be called on the parent to remove it.
 function LanguageTree:destroy()
   -- Cleanup here
+  if self._has_scratch_buf then
+    self._has_scratch_buf = false
+    vim.api.nvim_buf_delete(self._source, {})
+  end
   for _, child in pairs(self._children) do
     child:destroy()
   end
@@ -842,7 +856,7 @@ function LanguageTree:included_regions()
 end
 
 ---@param node TSNode
----@param source string|integer
+---@param source integer
 ---@param metadata vim.treesitter.query.TSMetadata
 ---@param include_children boolean
 ---@return Range6[]

src/nvim/lua/treesitter.c

@@ -513,24 +513,12 @@ static int parser_parse(lua_State *L)
     old_tree = ud ? ud->tree : NULL;
   }
 
-  TSTree *new_tree = NULL;
-  size_t len;
-  const char *str;
-  handle_T bufnr;
-  buf_T *buf;
-  TSInput input;
+  if (lua_type(L, 3) != LUA_TNUMBER) {
+    return luaL_argerror(L, 3, "expected buffer handle");
+  }
 
-  // This switch is necessary because of the behavior of lua_isstring, that
-  // consider numbers as strings...
-  switch (lua_type(L, 3)) {
-  case LUA_TSTRING:
-    str = lua_tolstring(L, 3, &len);
-    new_tree = ts_parser_parse_string(p, old_tree, str, (uint32_t)len);
-    break;
-
-  case LUA_TNUMBER:
-    bufnr = (handle_T)lua_tointeger(L, 3);
-    buf = handle_get_buffer(bufnr);
+  handle_T bufnr = (handle_T)lua_tointeger(L, 3);
+  buf_T *buf = handle_get_buffer(bufnr);
 
   if (!buf) {
 #define BUFSIZE 256
@@ -540,7 +528,9 @@ static int parser_parse(lua_State *L)
 #undef BUFSIZE
   }
 
-    input = (TSInput){ (void *)buf, input_cb, TSInputEncodingUTF8, NULL };
+  TSInput input = (TSInput){ (void *)buf, input_cb, TSInputEncodingUTF8, NULL };
+  TSTree *new_tree = NULL;
+
   if (!lua_isnil(L, 5)) {
     uint64_t timeout_ns = (uint64_t)lua_tointeger(L, 5);
     TSLuaParserCallbackPayload payload =
@@ -553,12 +543,6 @@ static int parser_parse(lua_State *L)
     new_tree = ts_parser_parse(p, old_tree, input);
   }
 
-    break;
-
-  default:
-    return luaL_argerror(L, 3, "expected either string or buffer handle");
-  }
-
   bool include_bytes = (lua_gettop(L) >= 4) && lua_toboolean(L, 4);
 
   if (!new_tree) {

test/functional/treesitter/parser_spec.lua

@@ -1264,6 +1264,12 @@ print()
       parser:for_each_tree(function(tstree, tree)
         ranges[tree:lang()] = { tstree:root():range(true) }
       end)
+
+      -- Scratch buffer should get cleaned up
+      assert(vim.api.nvim_buf_is_loaded(parser:source()))
+      parser:destroy()
+      assert(not vim.api.nvim_buf_is_loaded(parser:source()))
+
       return ranges
     end)