encoding: Update handling of encoding in tests

Always run tests with encoding=utf-8, regardless of user locale
Don't set &encoding after startup in tests

Helped-By: Michael Reed <m.reed@mykolab.com>

src/nvim/testdir/test10.in

@@ -4,7 +4,6 @@ STARTTEST
 :so small.vim
 :" Also test a BOM is ignored.
 :so mbyte.vim
-:set encoding=utf-8
 :7/start of errorfile/,/end of errorfile/w! Xerrorfile1
 :7/start of errorfile/,/end of errorfile/-1w! Xerrorfile2
 :/start of testfile/,/end of testfile/w! Xtestfile

src/nvim/testdir/test39.in

@@ -5,8 +5,6 @@ And test "U" in Visual mode, also on German sharp S.
 STARTTEST
 :so small.vim
 :so mbyte.vim
-:" This only works when 'encoding' is "latin1", don't depend on the environment
-:set enc=latin1
 /^abcde
 :" Test shift-right of a block
 jlllljj>wlljlll>
@@ -39,18 +37,18 @@ G$khhhhhkkcmno
 :exe ":norm! l\<C-V>j$hhAab\<Esc>"
 :.,/^$/w >> test.out
 :" Test for Visual block insert when virtualedit=all and utf-8 encoding
-:set ve=all enc=utf-8
+:set ve=all
 :/\t\tline
 :exe ":norm! 07l\<C-V>jjIx\<Esc>"
 :.,/^$/w >> test.out
 :" Test for Visual block append when virtualedit=all
 :exe ":norm! 012l\<C-v>jjAx\<Esc>"
-:set ve= enc=latin1
+:set ve=
 :.,/^$/w >> test.out
-:" gUe must uppercase a whole word, also when ß changes to SS
-Gothe youtußeuu endYpk0wgUe
+:" gUe must uppercase a whole word, also when ß changes to SS
+Gothe youtußeuu endYpk0wgUe
 :" gUfx must uppercase until x, inclusive.
-O- youßtußexu -0fogUfx
+O- youßtußexu -0fogUfx
 :" VU must uppercase a whole line
 YpkVU
 :" same, when it's the last line in the buffer
@@ -89,7 +87,7 @@ cccc
 dddd
 
 yaaa
-¿¿¿
+¿¿¿
 bbb
 
 A23

src/nvim/testdir/test69.in

@@ -5,7 +5,6 @@ Also test byteidx() and byteidxcomp()
 
 STARTTEST
 :so mbyte.vim
-:set encoding=utf-8
 ENDTEST
 
 Results of test69:

src/nvim/testdir/test83.in

@@ -2,7 +2,6 @@ Tests for tag search with !_TAG_FILE_ENCODING.
 
 STARTTEST
 :so mbyte.vim
-:set enc=utf8
 :if !has('iconv') || iconv("\x82\x60", "cp932", "utf-8") != "\uff21"
 : e! test.ok
 : w! test.out

src/nvim/testdir/test_eval.in

@@ -4,7 +4,6 @@ Note: system clipboard is saved, changed and restored.
 
 STARTTEST
 :so small.vim
-:set encoding=latin1
 :set noswapfile
 :lang C
 :fun AppendRegContents(reg)

src/nvim/testdir/unix.vim

@@ -1,3 +1,6 @@
 " Settings for test script execution
 " Always use "sh", don't use the value of "$SHELL".
 set shell=sh
+
+" Don't depend on system locale, always use utf-8
+set encoding=utf-8

test/functional/helpers.lua

@@ -7,7 +7,7 @@ local Session = require('nvim.session')
 
 local nvim_prog = os.getenv('NVIM_PROG') or 'build/bin/nvim'
 local nvim_argv = {nvim_prog, '-u', 'NONE', '-i', 'NONE', '-N',
-                   '--cmd', 'set shortmess+=I background=light noswapfile noautoindent laststatus=1',
+                   '--cmd', 'set shortmess+=I background=light noswapfile noautoindent laststatus=1 encoding=utf-8',
                    '--embed'}
 
 -- Formulate a path to the directory containing nvim.  We use this to

test/functional/legacy/044_099_regexp_multibyte_magic_spec.lua

@@ -33,7 +33,6 @@ local function run_test_with_regexpengine(regexpengine)
     j 0123❤x
     k combinations]])
 
-  execute('set encoding=utf-8')
   execute('set re=' .. regexpengine)
 
   -- Lines 1-8. Exercise regexp search with various magic settings. On each

test/functional/legacy/075_maparg_spec.lua

@@ -10,7 +10,6 @@ describe('maparg()', function()
 
   it('is working', function()
     execute('set cpo-=<')
-    execute('set encoding=utf8')
 
     -- Test maparg() with a string result
     execute('map foo<C-V> is<F4>foo')

test/functional/legacy/082_string_comparison_spec.lua

@@ -9,7 +9,6 @@ describe('case-insensitive string comparison in UTF-8', function()
   setup(clear)
 
   it('is working', function()
-    execute('set enc=utf8')
     feed('ggdG<cr>')
     source([[
       function! Ch(a, op, b, expected)

test/functional/legacy/095_regexp_multibyte_spec.lua

@@ -15,7 +15,7 @@ describe('regex with multi-byte', function()
       Results of test95:]])
 
     source([=[
-      set encoding=utf-8 nomore
+      set nomore
       let tl = []
 
       call add(tl, [2, '[[:alpha:][=a=]]\+', '879 aiaãâaiuvna ', 'aiaãâaiuvna'])

test/functional/legacy/mapping_spec.lua

@@ -12,8 +12,6 @@ describe('mapping', function()
       test starts here:
       ]])
 
-    execute('set encoding=utf-8')
-
     -- Abbreviations with р (0x80) should work.
     execute('inoreab чкпр   vim')
     feed('GAчкпр <esc>')

test/functional/legacy/utf8_spec.lua

@@ -10,7 +10,6 @@ describe('utf8', function()
   it('is working', function()
     insert('start:')
 
-    execute('set encoding=utf-8')
     execute('new')
     execute('call setline(1, ["aaa", "あああ", "bbb"])')