Add option setOrigRanges for parsing docs

[1nVitr0]

This pull request adds the following parameter to the document parser:

setOrigRanges?: boolean

It uses the same name as the corresponding function setOrigRanges, which sets original ranges on cstNodes for \r\n style line endings. Default is false and it obviously only has an effect when keepCstNodes is set to true, although it does get executed either way, just that the CST nodes get discarded when keepCstNodes is set to false.

I realize adding more options should always be considered carefully, but being able to set the original ranges when parsing a document was pretty useful for me.

[eemeli]

I'm going to hold off on really considering this until #203 gets merged, because that might remove the need for this: the new parser doesn't normalise the line endings. OTOH, atm it's only tracking the start indices of nodes, and I haven't figured out exactly yet if or how to provide end indices as well.

[eemeli]

Closing, as CRLF newlines are no longer converted to LF at any parsing stage, and range positions are always valid for the original string.