refactor(tests): remove redir_exec #15718

Problem
- `redir_exec` is obsolete, but it keeps getting used in new tests
  because people copy existing tests.
- Disadvantages of `redir_exec`:
  - Captures extra junk before the actual error/message that we _want_ to test.
  - Does not fail on error, unlike e.g. `command()`.

Solution
- Use new functions like `nvim_exec` and `pcall_err`.

1 files changed, 8 insertions, 2 deletions

diff --git a/test/functional/vimscript/execute_spec.lua b/test/functional/vimscript/execute_spec.lua
index fccf52935b..e21c71dc7f 100644
--- a/test/functional/vimscript/execute_spec.lua
+++ b/test/functional/vimscript/execute_spec.lua
@@ -3,7 +3,6 @@ local eq = helpers.eq
 local eval = helpers.eval
 local clear = helpers.clear
 local source = helpers.source
-local redir_exec = helpers.redir_exec
 local exc_exec = helpers.exc_exec
 local funcs = helpers.funcs
 local Screen = require('test.functional.ui.screen')
@@ -15,7 +14,14 @@ describe('execute()', function()
   before_each(clear)
 
   it('captures the same result as :redir', function()
-    eq(redir_exec('messages'), funcs.execute('messages'))
+    command([[
+      echomsg 'foo 1'
+      echomsg 'foo 2'
+      redir => g:__redir_output
+        silent! messages
+      redir END
+    ]])
+    eq(eval('g:__redir_output'), funcs.execute('messages'))
   end)
 
   it('captures the concatenated outputs of a List of commands', function()