Fix: Improve linting configuration and test runner

- Fix passing unused arguments with underscores
- Remove redefining of global print function
- Use sandboxed environments for test files
- Add missing Lua functions to globals list
- Update luacheck configuration for test files

Author: greggh

.luacheckrc

@@ -21,6 +21,10 @@ std = {
     "error",
     "assert",
     "_VERSION",
+    "loadfile",
+    "setfenv",
+    "setmetatable",
+    "getmetatable",
   },
 }
 
@@ -31,19 +35,53 @@ exclude_files = {
   "tests/plenary/*",
 }
 
+-- Special configuration for test files
+files["tests/**/*.lua"] = {
+  -- Allow common globals used in testing
+  globals = {
+    -- Common testing globals
+    "describe", "it", "before_each", "after_each", "teardown", "pending", "spy", "stub", "mock",
+    -- Lua standard utilities used in tests
+    "print", "dofile",
+    -- Test helpers
+    "test", "expect",
+    -- Global test state (allow modification)
+    "_G", 
+  },
+  
+  -- Allow modification of globals in test files
+  allow_defined_top = true,
+  
+  -- Allow mutation of underscore globals in test files
+  module = true,
+  
+  -- In tests, we're not concerned about these
+  max_cyclomatic_complexity = false,
+  
+  -- Don't report accessing/mutating globals that are needed for test frameworks
+  read_globals = {
+    "_G",      -- Global environment
+    "package", -- Package management
+  },
+  
+  -- Ignore unused variables in test code
+  unused = false,
+  unused_args = false,
+}
+
 -- Allow unused self arguments of methods
 self = false
 
 -- Don't report unused arguments/locals
 unused_args = false
 unused = false
 
--- Ignore warnings related to whitespace
+-- Ignore certain warnings that don't affect functionality
 ignore = {
   "611", -- Line contains trailing whitespace
   "612", -- Line contains trailing whitespace in a comment
   "613", -- Line contains trailing whitespace in a string
-  "614", -- Line contains only whitespace
+  "614", -- Line contains only whitespace (formatting style preference)
 }
 
 -- Maximum line length

tests/run_tests.lua

@@ -1,5 +1,5 @@
 -- Test runner for Plenary-based tests
-local ok, plenary = pcall(require, "plenary")
+local ok, _ = pcall(require, "plenary")
 if not ok then
   print("ERROR: Could not load plenary")
   vim.cmd("qa!")
@@ -26,7 +26,7 @@ _G.TEST_RESULTS = {
 
 -- Silence vim.notify during tests to prevent output pollution
 local original_notify = vim.notify
-vim.notify = function(msg, level, opts)
+vim.notify = function(msg, level, _)
   -- Capture the message for debugging but don't display it
   if level == vim.log.levels.ERROR then
     _G.TEST_RESULTS.last_error = msg
@@ -148,11 +148,10 @@ local function run_tests()
   end
 
   -- Setup a test counter based on terminal output
-  -- Override print to capture and count the test results
-  local old_print = print
-  print = function(msg)
+  -- Create an output processor function to count test results
+  local function process_output(msg)
     -- Forward to original print
-    old_print(msg)
+    print(msg)
 
     -- Count tests by looking for success/failure indicators
     if type(msg) == "string" then
@@ -171,19 +170,34 @@ local function run_tests()
     end
   end
 
-  -- Run each test file individually
+  -- We don't need to store the original print function
+  -- as we're using isolated environments for each test file
+
+  -- Run each test file individually in a protected environment
   for _, file in ipairs(test_files) do
-    print("\nRunning tests in: " .. vim.fn.fnamemodify(file, ":t"))
-    local status, err = pcall(dofile, file)
-    if not status then
-      print("Error loading test file: " .. err)
+    process_output("\nRunning tests in: " .. vim.fn.fnamemodify(file, ":t"))
+
+    -- Create an environment with a modified print function
+    local test_env = setmetatable({
+      print = process_output,
+    }, { __index = _G })
+
+    -- Load the file in this environment
+    local chunk, load_err = loadfile(file)
+    if not chunk then
+      process_output("Error loading test file: " .. load_err)
       error_counter = error_counter + 1
+    else
+      -- Set the environment and run the file
+      setfenv(chunk, test_env)
+      local status, err = pcall(chunk)
+      if not status then
+        process_output("Error executing test file: " .. err)
+        error_counter = error_counter + 1
+      end
     end
   end
 
-  -- Restore print function
-  print = old_print
-
   -- Count the actual number of tests based on output
   -- The test counts will be used for hardcoding since the dynamic counting isn't working
   local test_count = 0