define a text object as the boundary of two marks

Question

Is there a way to turn the "region bounded by two specific marks" into a text object similar to aw or ap?

I've been messing around with vi and found it convenient to use a and b. and define a region and to define commands that act on the region bounded by a and b.

It would be nice to be able to expose a text object for this purpose in vim.

---

I have the following bindings in vi for emulating something kind of like visual mode. I have some macros that set a and b to the boundaries of an interesting region. And another set of macros for manipulating the region bounded by a and b.

It's nice to be able to pick whether I want a linewise selection or not after setting the marks.

^A and ^P are literal control characters, in this case they set the mark a to the beginning of the buffer or paragraph and b to the end.

map ^A 1GmaGmb
map ^P }mb{jmak}

And other macros that do stuff to the region like indent it and dedent it.

map \. mp'a>'b`p
map \, mp'a<'b`p

Answer 1

Here is an example for a custom text object working with the marks a and b.

" char-wise 
onoremap am :<c-u>execute 'normal `av`b'<cr>
xnoremap am :<c-u>execute 'normal `av`b'<cr>
" line-wise
onoremap aM :<c-u>execute 'normal `aV`b'<cr>
xnoremap aM :<c-u>execute 'normal `aV`b'<cr>

To copy the characters between the marks a and b hit yam. To copy the lines between a and b hit yaM. You can use it with other commands., like =am to format the lines between marks a and b.

See :help :map-modes and :help omap-info.

Notes:

• This moves the cursor. After it hit Ctrl-O twice to get back to your initial cursor position.
• This changes the visual selection. If you used yam and hit gv afterwards, the visual selection is mark a to b.

---

If you want to work with any mark, you could try this:

onoremap <silent> am :<c-u>execute 'normal `'. nr2char(getchar()) . 'v`' . nr2char(getchar())<cr>
xnoremap <silent> am :<c-u>execute 'normal `'. nr2char(getchar()) . 'v`' . nr2char(getchar())<cr>
onoremap <silent> aM :<c-u>execute 'normal `'. nr2char(getchar()) . 'V`' . nr2char(getchar())<cr>
xnoremap <silent> aM :<c-u>execute 'normal `'. nr2char(getchar()) . 'V`' . nr2char(getchar())<cr>

Now you need to give the marks. Like yamab copies the range of the marks a and b, while yamxy uses the marks x and y. If you only want the line of the mark a copied, you could use yaMaa.

Be careful with upper case marks here. Upper case marks might change the buffer. Assuming mark A is in another buffer, the command yamAb will switch to the buffer of the mark A and will copy to mark b in that buffer.

Answer 2

I would use something like the following. There might be a lot of corner cases, but it's a start:

for s:nr1 in range(97, 122)
    for s:nr2 in range(97, 122)
        let s:char1 = nr2char(s:nr1)
        let s:char2 = nr2char(s:nr2)
        execute 'xnoremap i'.s:char1.s:char2.' :<C-u>normal! `'.s:char1.'v`'.s:char2.'<CR>'
        execute 'onoremap i'.s:char1.s:char2.' :normal vi'.s:char1.s:char2.'<CR>'
    endfor
endfor

I used range(97, 122) to get a list of numbers from 97 to 122 (see :h range() and these represent the ASCII values of characters a to z which could be converted to characters with :h nr2char(). So, if you want uppercase letter marks, you can edit the list range yourself.