D4997: absorb: use a formatter to generate output

[Yuya Nishihara]

> +        fm.startitem()
> +        fm.write('hunk', '        %s\n',
> +                 '@@ -%d,%d +%d,%d @@'
> +                 % (a1, a2 - a1, b1, b2 - b1), label='diff.hunk')
> +        fm.data(path=self.path, linetype='hunk')
> +
> +        def writeline(idx, diffchar, line, linetype, linelabel):
> +            fm.startitem()
> +            node = ''
> +            if idx:
> +                ctx = self.fctxs[idx]
> +                fm.context(fctx=ctx)
> +                node = ctx.hex()
> +            fm.write('node', '%-7.7s ', node, label='absorb.node')
> +            fm.write('diffchar ' + linetype, '%s%s\n', diffchar, line,
> +                     label=linelabel)

Nit: maybe rename `"diffchar"` to `"change"`? It's called as such in "hg grep".
https://www.mercurial-scm.org/wiki/GenericTemplatingPlan#Dictionary

> +            fm.data(path=self.path, linetype=linetype)
> +
> +        for i in pycompat.xrange(a1, a2):
> +            writeline(aidxs[i - a1], '-', trim(alines[i]), 'deleted',
> +                      'diff.deleted')
> +        for i in pycompat.xrange(b1, b2):
> +            writeline(bidxs[i - b1], '+', trim(blines[i]), 'inserted',
> +                      'diff.inserted')

Perhaps, it's better to make deleted/inserted lines nested to preserve the
original data structure.

```
{"hunk", ...,
 "deletedlines": [{"line": ...}, ...]
 "insertedlines": ...}
```

or

```
{"hunk", ...,
 "lines": [{"line": ..., "deleted": True}, ...]}
```

instead of

```
{"hunk", ...},
{"deleted": ...},
{"inserted": ...},
```