perf(tags): improve parsing logic (#2657)

Author: ibhagwan

lua/fzf-lua/actions.lua

@@ -268,8 +268,11 @@ M.vimcmd_entry = function(vimcmd, selected, opts, bufedit)
         pcall(utils.jump_to_location, { uri = entry.uri, range = entry.range }, "utf-16",
           opts.reuse_win)
       elseif entry.line == 0 and entry.ctag then
-        vim.api.nvim_win_set_cursor(0, { 1, 0 })
-        vim.fn.search(entry.ctag, "W")
+        local re = utils.ctag_to_magic(entry.ctag)
+        if utils.vim_regex(re, opts) then
+          vim.api.nvim_win_set_cursor(0, { 1, 0 })
+          vim.fn.search(re, "W")
+        end
       elseif not opts.no_action_set_cursor and entry.line > 0 or entry.col > 0 then
         -- Make sure we have valid line/column
         -- e.g. qf lists from files (no line/col), dap_breakpoints

lua/fzf-lua/defaults.lua

@@ -703,32 +703,19 @@ M.file = function(x, opts)
 end
 
 M.tag = function(x, opts)
-  local name, file, text = x:match("([^\t]+)\t([^\t]+)\t(.*)")
-  if not file or not name or not text then return x end
-  text = text:match([[(.*);"]]) or text -- remove ctag comments
-  -- unescape ctags special chars
-  -- '\/' -> '/'
-  -- '\\' -> '\'
-  for _, s in ipairs({ "/", "\\" }) do
-    text = text:gsub([[\]] .. s, s)
-  end
+  local name, file, excmd = x:match("([^\t]+)\t([^\t]+)\t(.*)")
+  if not file or not name or not excmd then return x end
+  local line, tag = path.parse_ctag_excmd(excmd, true)
   -- different alignment fmt if string contains ansi coloring
   -- from rg/grep output when using `tags_grep_xxx`
-  local align = utils.has_ansi_coloring(name) and 47 or 30
-  local line, tag = text:match("(%d-);?(/.*/)")
-  if not tag then
-    -- lines with a tag located solely by line number contain nothing but the
-    -- number at this point (e.g. using "ctags -R --excmd=number")
-    line = text:match("%d+")
-  end
-  line = line and #line > 0 and tonumber(line)
-  return string.format("%-" .. tostring(align) .. "s%s%s%s: %s",
+  local align = 24
+  if utils.has_ansi_coloring(name) then align = align + 17 end
+  return string.format("%-" .. tostring(align) .. "s\t%s\t%s%s",
     name,
-    utils.nbsp,
     M.file(file, opts),
-    not line and "" or ":" .. utils.ansi_codes.green(tostring(line)),
-    utils.ansi_codes.blue(tag)
-  ), line
+    line and utils.ansi_codes.green(tostring(line)) .. ";" or "",
+    tag and utils.ansi_codes.blue(tag) or ""
+  )
 end
 
 M.git_status = function(x, opts)

lua/fzf-lua/make_entry.lua

@@ -377,21 +377,43 @@ function M.lengthen(path)
       or string.format("<glob expand failed for '%s'>", glob_expr)
 end
 
-function M.entry_to_ctag(entry, noesc)
-  local ctag = entry:match("%:.-[/\\]^?\t?(.*)[/\\]")
-  -- if tag name contains a slash we could
-  -- have the wrong match, most tags start
-  -- with ^ so try to match based on that
-  ctag = ctag and ctag:match("[/\\]^(.*)") or ctag
-  if ctag and not noesc then
-    -- required escapes for vim.fn.search()
-    -- \ ] ~ *
-    ctag = ctag:gsub("[\\%]~*]",
-      function(x)
-        return "\\" .. x
-      end)
-  end
-  return ctag
+---@param excmd string
+---@param do_not_strip_slashes boolean?
+---@return number?, string?
+function M.parse_ctag_excmd(excmd, do_not_strip_slashes)
+  -- remove anything after excmd ;" postfix
+  excmd = excmd:gsub([[;".-$]], ";")
+  -- test for "--excmd={number|combine}"
+  local line = excmd:match("^(%d+);")
+  -- adjust for line, position cursor after ;
+  if line then excmd = excmd:sub(#line + 2) end
+  -- extract ctag
+  local tag = do_not_strip_slashes
+      and excmd:match("/.*/")
+      or excmd:match("/(.*)/")
+  return tonumber(line), tag
+end
+
+---@param raw string
+---@param opts table
+---@return fzf-lua.path.Entry
+function M.entry_to_ctag(raw, opts)
+  assert(opts._ctag)
+  local _, file, excmd = raw:match("([^\t]+)\t([^\t]+)\t(.*)")
+  if not file or not excmd then return {} end
+  file = file:match(".*" .. utils.nbsp .. "(.+)$") or file
+  local cwd = opts.cwd or opts._cwd
+  if cwd and not M.is_absolute(file) then file = M.join({ cwd, file }) end
+  if opts.path_shorten then file = M.lengthen(file) end
+  local line, tag = M.parse_ctag_excmd(excmd)
+  return {
+    path = file,
+    ctag = tag,
+    line = line or 0,
+    col = 0,
+    stripped = string.format("%s:%s %s", file, line and line .. ":" or "", tag or ""),
+    debug = opts.debug and raw:match("^%[DEBUG]") and raw or nil,
+  } ---@as fzf-lua.path.Entry
 end
 
 ---@param entry string
@@ -437,6 +459,7 @@ end
 function M.entry_to_file(entry, opts, force_uri)
   -- NOTE: see note in meta.lua:global regarding alt options
   opts = opts and opts.__alt_opts or opts or {}
+  assert(opts)
   if opts._fmt then
     if type(opts._fmt._from) == "function" then
       entry = opts._fmt._from(entry, opts)
@@ -450,6 +473,10 @@ function M.entry_to_file(entry, opts, force_uri)
   if opts.render_crlf then
     entry = entry:gsub("␊", "\n"):gsub("␍", "\r")
   end
+  -- ctag processing is done separately
+  if opts._ctag then
+    return M.entry_to_ctag(entry, opts)
+  end
   local stripped, idx = (function()
     -- Returns the first viable path:line?:col? + rest of line
     -- stripping until the last occurrence of utils.nbsp may err
@@ -556,7 +583,6 @@ function M.entry_to_file(entry, opts, force_uri)
     line     = utils.tointeger(type(opts.line_query) == "function" and
       (opts.line_query(FzfLua.get_info().query)) or line) or 0,
     col      = utils.tointeger(col) or 0,
-    ctag     = opts._ctag and M.entry_to_ctag(stripped) or nil,
     debug    = opts.debug and entry:match("^%[DEBUG]") and entry or nil,
   }
 end

lua/fzf-lua/path.lua

@@ -1453,6 +1453,7 @@ function Previewer.buffer_or_file:set_cursor_hl(entry)
 
   -- If called from tags previewer, can happen when using ctags cmd
   -- "ctags -R --c++-kinds=+p --fields=+iaS --extras=+q --excmd=combine"
+  -- vim.regex is always magic, see `:help vim.regex`
   regex = regex and #regex > 0 and utils.regex_to_magic(regex)
       or entry.ctag and utils.ctag_to_magic(entry.ctag)
 
@@ -1481,20 +1482,14 @@ function Previewer.buffer_or_file:set_cursor_hl(entry)
     -- If regex is available (grep/lgrep), match on current line
     if regex and hls.search then
       local regex_start, regex_end
-      -- vim.regex is always magic, see `:help vim.regex`
-      ---@diagnostic disable-next-line: param-type-mismatch
-      local reg = vim.F.npcall(vim.regex, regex)
+      local reg = utils.vim_regex(regex, { silent = true })
       if reg then
         if regex ~= regex:lower() then
           regex_start, regex_end = reg:match_line(buf, lnum - 1, col - 1)
         else
           local line = api.nvim_buf_get_lines(buf, lnum - 1, lnum, false)[1] or ""
           regex_start, regex_end = reg:match_str(line:sub(col):lower())
         end
-      elseif self.opts.silent ~= true then
-        utils.warn(
-          [[Unable to init vim.regex with "%s", %s. . Add 'silent=true' to hide this message.]],
-          regex, reg)
       end
       if regex_start and regex_end then
         extmark = api.nvim_buf_set_extmark(buf, self.ns_previewer, lnum - 1, regex_start + col - 1, {
@@ -1766,7 +1761,9 @@ function Previewer.tags:set_cursor_hl(entry)
     -- didn't reload the buffer (same file)
     api.nvim_win_set_cursor(0, { 1, 0 })
     fn.clearmatches()
-    local ctag = assert(entry.ctag)
+    local ctag = utils.ctag_to_magic(assert(entry.ctag))
+    -- test the regex so we can alert the user of the search fail
+    if not utils.vim_regex(ctag, self.opts) then return end
     fn.search(ctag, "W")
     if self.win.hls.search then
       fn.matchadd(self.win.hls.search, ctag)

lua/fzf-lua/previewer/builtin.lua

@@ -182,34 +182,26 @@ end
 
 local grep_tag = function(file, tag)
   local line = 1
-  local filepath = file
-  local pattern = utils.rg_escape(vim.trim(tag))
-  if not pattern or not filepath then return line end
-  local grep_cmd = vim.fn.executable("rg") == 1
-      and { "rg", "--line-number" }
-      or { "grep", "-n", "-P" }
+  if not tag or not file then return line end
+  local pattern = utils.ctag_escape(tag)
   -- ctags uses '$' at the end of short patterns
   -- 'rg|grep' does not match these properly when
   -- 'fileformat' isn't set to 'unix', when set to
   -- 'dos' we need to prepend '$' with '\r$' with 'rg'
-  -- it is simpler to just ignore it completely.
-  --[[ local ff = fileformat(filepath)
-  if ff == 'dos' then
-    pattern = pattern:gsub("\\%$$", "\\r%$")
-  else
-    pattern = pattern:gsub("\\%$$", "%$")
-  end --]]
   -- equivalent pattern to `rg --crlf`
   -- see discussion in #219
-  pattern = pattern:gsub("\\%$$", "\\r??%$")
+  pattern = pattern:gsub("%$$", "\\r??%$")
+  local grep_cmd = vim.fn.executable("rg") == 1
+      and { "rg", "--line-number" }
+      or { "grep", "-n", "-P" }
   local cmd = utils.tbl_deep_clone(grep_cmd)
   table.insert(cmd, pattern)
-  table.insert(cmd, filepath)
+  table.insert(cmd, file)
   local out, rc = utils.io_system(cmd)
   if rc == 0 then
     line = utils.tointeger(out:match("[^:]+")) or 1
   else
-    utils.warn(("previewer: unable to find pattern '%s' in file '%s'"):format(pattern, file))
+    utils.warn("Unable to match pattern '%s' in file '%s'", pattern, file)
   end
   return line
 end
@@ -228,16 +220,9 @@ function Previewer.cmd_async:parse_entry_and_verify(entrystr)
   -- make relative for bat's header display
   local filepath = path.relative_to(entry.bufname or entry.path or "", utils.cwd())
   if self.opts._ctag then
-    -- NOTE: override `entry.ctag` with the unescaped version
-    entry.ctag = path.entry_to_ctag(entry.stripped, true)
-    if not tonumber(entry.line) or tonumber(entry.line) < 1 then
+    if not entry.line or tonumber(entry.line) < 1 then
       -- default tags are without line numbers
-      -- make sure we don't already have line #
-      -- (in the case the line no. is actually 1)
-      local line = entry.stripped:match("[^:]+(%d+):")
-      if not line and entry.ctag then
-        entry.line = grep_tag(filepath, entry.ctag)
-      end
+      entry.line = grep_tag(filepath, entry.ctag)
     end
   end
   local errcmd = nil
@@ -316,7 +301,7 @@ function Previewer.bat_async:cmdline(o)
     local filepath, entry, errcmd = self:parse_entry_and_verify(items[1])
     if not filepath or not entry then return utils.shell_nop() end
     local line_range = ""
-    if entry.ctag and entry.line then
+    if self.opts._ctag and entry.line then
       -- this is a ctag without line numbers, since we can't
       -- provide the preview file offset to fzf via the field
       -- index expression we use '--line-range' instead

lua/fzf-lua/previewer/fzf.lua

@@ -201,8 +201,8 @@ M.btags = function(opts)
   -- Used as fallback to pipe the tags into fzf from stdout
   opts._btags_cmd = string.format("%s %s %s",
     opts.ctags_bin or "ctags",
-    opts.ctags_args or "-f -",
-    opts.filename)
+    opts.ctags_args or "-f - --excmd=combine",
+    libuv.shellescape(opts.filename))
   if opts.ctags_autogen then
     opts.cmd = opts.cmd or opts._btags_cmd
   end

lua/fzf-lua/providers/tags.lua

@@ -216,6 +216,20 @@ function M.is_darwin()
   return uv.os_uname().sysname == "Darwin"
 end
 
+---@param re string
+---@param opts table
+---@return vim.regex?
+function M.vim_regex(re, opts)
+  local ok, regex = pcall(vim.regex, re)
+  if ok then return regex --[[@as vim.regex]] end
+  if not ok and (not opts or opts.silent ~= true) then
+    M.warn(
+      [[Unable to init vim.regex with "%s", %s. . Add 'silent=true' to hide this message.]],
+      re, regex)
+    return nil
+  end
+end
+
 ---@param str string
 ---@return string
 function M.rg_escape(str)
@@ -230,16 +244,34 @@ function M.rg_escape(str)
   return ret
 end
 
+---@param str string
+---@return string
 function M.regex_to_magic(str)
   -- Convert regex to "very magic" pattern, basically a regex
   -- with special meaning for "%=&<>~", `:help /magic`
-  return [[\v]] .. str:gsub("[%%=&@<>~]", function(x)
-    return "\\" .. x
-  end)
+  return [[\v]] .. str:gsub("([%%=&<>])", [[\%1]])
+      -- searching for @ in very magic needs [@]
+      :gsub("([@])", "[%1]")
 end
 
+---@param str string
+---@return string
+function M.ctag_escape(str)
+  -- unescape already escaped ctags slashes
+  -- '\\' -> '\'
+  -- '\/' -> '/'
+  str = str:gsub([[\\]], [[\]])
+  str = str:gsub([[\/]], [[/]])
+  -- regex escape
+  return (M.rg_escape(str)
+    -- unescape ^$ if were positioned in start/end respectively
+    :gsub([[^\^]], "^"):gsub([[\$$]], "$"))
+end
+
+---@param str string
+---@return string
 function M.ctag_to_magic(str)
-  return [[\v]] .. str:gsub("[=&@<>{%(%)%.%[]", function(x) return [[\]] .. x end)
+  return M.regex_to_magic(M.ctag_escape(str))
 end
 
 function M.sk_escape(str)