Excellent vim doc:
  http://www.ukuug.org/events/linux2004/programme/paper-SMyers/Linux_2004_slides/vim_tips/

help
    :help keyword CTRL-D
    The CTRL-D shows completion matches.
    The help system uses tags, so see TAGS below.

Manual notation
    <C-..., <S-... <M-... or <A-...  mean CTRL + or SHIFT + or META/ALT +.
    CTRL-D  obvious

e object is just like w object, except it doesn't include the trailing space

% means same as 1,$.  E.g:  :%s/X/Y/ switches X to Y on all lines.  AWESOME!

switch/replace  replacement string may contain \n backreferences like \1, etc.

CTRL-R undoes previous undos

U is u for entire line

c$ changes to $ (like cw)

% goes to the matching (), [] or {}.

O opens a line Above the current line

To turn on/off search highlighting
:set hls
:set nohls

To turn off auto-move during search
:set is
:set nois

    1. one
    2. two

UTF-8 = Unicode
ISO-8859-1 = Western


FOREIGN CHARACTERS
    Text files with these are ISO-8859 files.
     (If the file type shows as "Non-ISO extended ASCII" it has other binaries).
    :digraphs     lists all available digraphs.
                  Remember to set encoding to utf8 if you want the full set.
                  Numbers shown are decimal codes, which is same as first
                    number shown by 'ga' command.

    The following hi-bit characters.  (Any char >= \x80 is high-bit).
               utf8       digraph keys (first + : + second)
           -----------    ¯-
        ä   \0344 \xe4    a:
        Ä   \0304 \xc4    ...
        ö   \0366 \xf6    ...
        Ö   \0326 \xd6    ...
        ü   \0374 \xfc    ...
        Ü   \0334 \xdc    ...
        ß   \0337 \xdf    ...
        «   \0253 \xab    <<
        »   \0273 \xbb    >>
        °                 DG
    Vim digraphs means that you enter two keystrokes to make one character.
     (:set digraph)
    The character itself may be 1 or 2 bytes (1 for latin1 and 2 for extended
     utf-8 characters).
     Like "Ctrl-K a:" or "a<BS>:" --> umlauted "a".
     (the 1-char output may be 1 or 2 bytes).
    UTF-8:  Special characters stored in 2 bytes.
    Remember, digraph is just Vim's very convenient method to enter special
     characters.  Encoding (utf-8 vs. latin1, etc.) is set via
     "set encoding=..." or env var. LANG, *LC*, etc.
    RedHat screws up foreign characters.
        With some terminal emulators, you can not see the foreign characters 
        no matter what.
        When you can see them, digraph sequences create 2-character outputs.
        They show file type of "UTF-8 Unicode English text".
            ä 228 = \0xe4 \0344  +
            Ä \0304 +
            ü \0374 +
            Ü \0334 +
            ö \0366 +
            Ö \0326 +
            « \0253 +
            » \0273 +
            ß \0337 +
    Digraphs are NOT supposed to be 2-character sequences!  (I think that
     this depends on encoding).  WRONG.
    Control mode (not ex mode) command "ga" tells decimal/hex/octal value/code 
     of character currently under the cursor.
    ISO-8859/UTF-8 Unicode (2-char "digraph" sequences)
        Digraphs work good (:help dig), but don't show up with some terminal 
        emulators.
        With "set digraph":
        (Seems that files with digraphs are called "UTF-8 Unicode".  (I seem
         to recall them being identified as ISO-8859 in some contexts though).
        Digraphs are 2-character sequences.
        German letter digraphs all begin with 0xc3.
          (e.g. ö = \0xc3 \0x96 = \0303 \0226)
        But « and » begin with 0xc2.
        a<BS>:, etc.
        s<BS>s

    Extended ASCII (1-chars)
    :help i_CTRL-v_digit
    <CTRL>-v + {
        Control or Alt sequence
        OCTAL
        xHEX
        u4CHARHEX
        U8CHARHEX
    }

VIM encoding values
latin1 == ISO-8859-1
utf-8 == UTF-8

Can set character DISPLAY encoding with :set encoding=X
    ! Generally DO NOT CHANGE THIS !
    This just needs to satisfy your terminal emulator; changing it can hose yor file content!
Can set character WRITING encoding with :set fileencoding=X
    Defaults according to LANG variable or OS or setting in .vimrc.
    Otherwise uses value from env var $LANG (which is UTF-8 w/ SUSE).
       (vim translates env vars to lower-case).
    Can convert a file's encoding by simply changing 'fileencoding' and writing again.


FILES always written according to 'fileencoding'.  If that not set then
according to 'encoding' (regardless whether file new or not).

THINGS GET FUCKED UP if you change 'encoding' with a file loaded.
It actually changes the file bytes and the file will not get written according
to 'fileencoding'

EXISTING FILES get 'fileencoding' set automatically according-to-the-file
when file is loaded.
AND 'encoding' gets set according to '--cmd encoding=x' OR LANG=x vim... or OS.
But I don't think there's any advantage to ever changing this from utf-8, since
that will handle characters in all of my files (if start using 4 byte characters
then consider utils utf-16).


:digraphs     lists all available digraphs.


To set indent size of 2 for XML files, add

    set shiftwidth=2 ts=2

to the bottom of VIMHOME/indent/xml.vim.
SEEMS THEY HAVE FIXED THE CONFIG FILE RUNTIMEPATH NOW!
See your runtime path with ":set rtp".  Should list $HOME/.vim up top.
Put your filetype-specific indent settings in $HOME/.vim/indent/EXTENSION.vim.
Like
    set sw=2 ts=2


To stop gvim from starting up with its own bad estimation of the window
width you need, use .vimrc/.gvimrc setting "set columns=80".

gvim == vim -g

Can use "echo 'x'" in .vim files.  Will display a popup or write to
status line, depending on current mode.

:set rtp
to echo the runtime path.  This is where you need to create subdirectories
ftpdetect and ftplugin.  Put filetype-handlers in ftplugin, like
vimfiles/ftplugin/text.vim to hand *.text files.

I have a note with subject "great putty .vimrc command" that says:
hi Normal ctermfg=white

ENDCODING conversion:
    iconv -f iso-8859-1 -t utf-8 -o new.utf8  # writes stdout by default
    # Does not do in-place conversions
The Suse release notes imply that ISO8859-1 is legacy.  I guess UTF8 is
what you should use if you need internationality.

INSERTION COMPLETION
  :help ins-completion
CTRL-N tries to auto-complete with something in current doc or tags.
    It is not fs-type-specific.
    CTRL-P is supposed to search in opposite direction, but I see no difference.
    (CTRL-X CTRL-N is very similar but just looks in current file).
CTRL-X Ctrl-O   OMNI-Completion.  FS-type-specific.
CTRL-X   lists ins-completion types available
    Ctrl-E Escapes out.
TABBIFYING COMPLETIONS
    Throw supertab.vim into $HOME/.vim/plugin directory.
    Says it does CTRL-P by default, but it is doing CTRL-N + CTRL-P.
    See "Global Variables" in supertab.vim for how to change.
    Can change at runtime with ":SuperTabHelp".
    With my ~/.vimrc additions, seems to work with <CTRL>+p

TAGS.  Keywords that you can "drill down".

    The source location is colored Magenta.
    Identifying keywords.
        Cyan (with my setup).
        Green (with my setup)
        Some blue words???
    These are for tagged "words".  Won't work unless tags have been generated
     for your doc using "ctags" or similar program (has been done for vim help).
    FORWARD
        ":tag word"
        Ctrl-]
        double<LeftMouse>
        <C-LeftMouse>
    RETURN
        ":pop"
        CTRL-T
        <C-RightMouse>
        (Or, CTRL-O more generally Jumps back)
    ":tags" shows current tag stack

SHIFT-click.  Highlights the clicked work in all occurrences.

FS TRAVERSAL
Just open a directory.
    Enter on an entry to open a file or directory.
    CTRL-O to back out.
    CTRL-Z CTRL-F to ins-completion a filepath.

SPELL CHECKING
    While in insert mode:  Ctrl-X s
    Must turn on PER DOCUMENT with ":setlocal spell spelllang=en_us"

eclim project may be for integration of vim with Eclipse.

CHANGING CASE/Capitalization
    Operation:  U = Upper; u = lower; ~ = invert
    Scoper:  repeat Op. is line; w is word; G to-end-of-doc; j3 to-3-lines-down; k4 to-4-lines-up
    guu   current line
    gUU   current line
    g~~   current line
    gu<scoper>
    gU<scoper>

FOLDING = z* commands.
Nomenclature:  fold vs. open
    Folding (vim) = Closing = Collapsing (general),
                        thus fold = close = collapse
    Open = expand = unfold
Levels are global.  If I am at global level 2 anywhere and do zm, it will fold
elements with content to level 2 globally.
    zfap:  create a Paragraph fold
    zo: Open (expand) current fold
    zc: FOLD = Close (collapse) current fold
    zr = zo *:  Reduce Folding at current level
    zm = zc *:  More Folding at current level
    zR = zo *:  Reduce all Folding
    zM = zc *:  More all Folding
    zi:  Toggle folding
foldcolumn's are EXTRA columns to the left to indicate folding.  Defaults to
0.
In graphical mode, the indicators can be clicked to fold/open.
IMPORTANT:  Must increase winwidth to accommodate the extra columns!

XML
    http://www.pinkjuice.com/howto/vimxml/

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
FUCKING $HOME/.vim/after/*/* files DON'T FUCKING WORK HALF THE TIME!!!!!!!!!
They are executed, but have no effect on ft-specific settings.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Seems this problem may have to do with "autocmd" commands.  My text.win
commands to set textwidth to 80 have NO EFFECT.  But if I put
    autocmd FileType  text setlocal textwidth=80
in plain .vimrc, it just works.

WRAPPING
"textwidth" setting controls what is the more important aspect of wrapping--
 automatic insertion of line breaks into the data.
The "wrap" option just controls whether or not to
 visually-wrap-vs-visually-truncate lone lines in the vim window.
N.b. vim defaults textwidth to 78 for text files (with local setting via
 autocmd).  This is a really bad setting IMO, since you are changing data to
  accommodate presentation conditions which are usually not present.
 Better to size lines to 80, which works directly with most use cases.
 It's pretty rare to need "nowrap" mode or foldcolumn's when you can't resize
  the Windows.  Even a remote terminal emulator can resize vim text windows.

gvimdiff colors:
    Real changes are in
       Light gray-blue (not plain light-blue)
       Red

ga shows character info

Great reference:  http://www.pixelbeat.org/vim.tips.html


EASY MODE
Looks like best for the ignorant is "egvim" == "vim -y -g".
How set easy Y mode from .vimrc file?

VIM is used to override ~ for user-specific config files.
VIMRUNTIME is obviously for the shared distro files.

Backup files.
   Disable with:  :nobackup  (says may also need :nobackupwrite or similar)
   :set backupdir=~/path1
   :set backupdir+=/dir2
   :set backup
   :set backupext=.bak
   Beware of :backupskip

Encryption:
Encrypt with command:  :X   and then CHANGE FILE! and save (regularly).
                                  OR write to a new file with :w newfile.txt
If you don't save the file then it won't really save at all (even with !).
Save decrypted with command: :set key=   (or :X with empty str) and then CHANGE FILE and save.
                                          OR write to a new file
Damned good thing:  Encrypting with empty string is same as encrypting with nothing.
  Therefore if you accidentally do ":X" just give it empty string twice and nothing will encrypt.

Changing capitalization/upper-case/lower-case
    ~ operator toggles
    U upper-cases
    u lower-cases
    g[Uu~]<MOVEMENT>
    g[Uu~][Uu~] or V[Uu~] does it to current line
    ggg[Uu~]G  does it to entire file

Search for non-ASCII character
DON'T WORK:
    yl  (that is an L)
    /<Ctrl-R><Ctrl-R>
ga  tells what the character is
Then search using hex:  /\%x1b
  or with Ctrl-V + sequence  like /<Ctr-v>@  to search for \0

To toggle line-numbering:
  :set number|nonumber

For better regexp goodness, extended patterns/regular express, need to prefix
regexp with \v for "very magic".
Quick is-very-magic-active test:  search for "<x" (any character x) like: /<x
    Regular mode:  It hits on literal <x
    Very magic:  It hits on words beginning with x
STILL NOT NORMAL regexp though! because with regular regexp < is a normal
character but vim very magic does this
    "every following character except a-zA-Z0-9_ special"
E.g. <word means beginning of word; word> means at end.  \b still does not work!
[To search for any character other than \w characters, you must search for \X ]
   ???  Search for \x\W seems to work great.  ???
There is no way to make this the default, so need hacks documented at
http://vim.wikia.com/wiki/Simplifying_regular_expressions_using_magic_and_no-magic
nnoremap / /\v
vnoremap / /\v
cnoremap %s/ %smagic/
cnoremap \>s/ \>smagic/
nnoremap :g/ :g/\v
nnoremap :g// :g//

File suffix associations in Windows.
At least with Windows 10, can't use UI for this any more.
assoc .txt
ftype txtfile="C:\path to gvim*.exe" "%1"


SWAP/BACKUP/UNDO Files
To prevent tons of temp files all over the place like weeds, use dedicated
directories.  The undo dir valid since vim v.7.3.
    mkdir -p ~/vim.d/backup.d ~/vim.d/undo.d ~/vim.d/swap.d
    set backupdir=~/vim.d/backup.d//
    set undodir=~/vim.d/undo.d//
    set directory=~/vim.d/swap.d//
The final // says to translate "/" in paths to %.
Append ",/dir/ectory" to add backup directories, e.g. ",c:/Windows/Temp//"

nmaps like so.  N.b. ex commands required a line break after.  Compound:
You type the "=j" in regular non-insert mode, not in ex mode.
nmap =j :%!python -m json.tool
:%s/  / /g

gvim Edit > Select All doesn't work well, so memorize:  ggVG

Allow to scroll cursor (not just view) to very beginning or end of document:
set scrolloff=0   

Select-all by gvim pulldown menu usually takes me from 2 to 6 attempts.
Fuck that.  Just use key sequence ggVG.

Entering characters by character number, like for binary characters.
CTRL+V NNN   decimal
CTRL+V oNNN  (from oN)  hex
CTRL+V xNN   hex
CTRL+V uNNNN hex

To persist font setting create or add to a ~/.gvimrc with something like:
set guifont=Lucida_console:h16

diff mode without -o (or is it -O?) splits into L & R panels.
In non-graphical mode you can move the vertical divider position with: vertical resize 40
(with CTRL+w focus in the panel to resize).

To sort contents with or without uniquifying:
:sort
:%sort u