[search] nimgrep option --machine_readable to allow integrating with other tools

[timotheecour]

I'd like to have an nimgrep option --machine_readable to be easily parsable by a program, eg:

file:line:col:nb_bytes_match:match:nb_bytes_line:line

which would print results 1 per line with all available context, eg:

lib/system.nim:1586:3:4:repr:100:*[T](x: T): string {.magic: "Repr", noSideEffect.}

currently it's tricky to use (and filter results via grep) because the context (file) is not shown in each line, eg:

lib/system.nim
  1586: proc repr*[T](x: T): string {.magic: "Repr", noSideEffect.}
  1586: proc repr*[T](x: T): string {.magic: "Repr", noSideEffect.}

so it's hard to filter via cmd line eg grep.

this would enable something like:

nimgrep -w --machine_readable --ignoreStyle "$1" **/*.nim| grep -E '\b(iterator|proc|template|macro) '

links

#7419

[Araq]

We have nimsuggest for editor integration.

[Araq]

This was implemented with --oneline.