Add sorting, faster demo, primitive initFileArray type check

Add sorting via command-line via nio order & nio emerge. These also show how to use the new x.save and var y = initFileArray[..](..) APIs.

Fix a couple bad bugs in nio cut and minor formatting stuff in nio moments.

Add a simple, faster demo that is basically a single-core memory BW benchmark (that also shows how to use initFileArray..). Also add a stays-in binary table generator.

Also add some guards to allow bypass of bounds checking in -d:danger mode.

Add `[]`, `save`, `FileArray`; Lessen warnings

Make some simple use cases simpler to use. Now you can just run someSeq.save and get back a slightly less capable version via initFileArray and access any NFile (read/write) with [index] that gives you a pointer in the general row format case (since only external type data is available).

See git commit log & code for more details.

Add `-d` to `fromSV`

Add a very simple feature to ease some use cases.

Various improvements based on using..

Probably most significant is the ability to share a string repo across some select set of columns.

Another is "@@" referring to the globally shared string repo, not "@." so that "@../../" parent directory is allowed to share strings across parses.

Somewhat better logging of errors, but this still needs a bit of work.

Add catz, nmatch, auto-schemas

catz is a one-stop shopping way to decode/decompress many common formats.

nmatch is useful for analyzing file formats - it simply prefixes a line with the number of pattern matches within it.

nio inferT can now generate a (provisional) parsing schema for you based on parsability within some representative input file, if you have one. Also, strip leading & trailing whitespace while parsing.