Better ACF pValues and use pVal not 1-pVal

Add the new spfun/studentT.ccPv for exact pValues(autocorrs) rather than approximations that sometimes do not even converge asymptotically. Since this can get slow for many thousands of points, keep old formula as a default for now.

Version bump just to update nimble dep on cligen-1.7.0

To be sure the fmtUncertain* bug fix propagates.

'nuff said.

0.4.1

A couple minor fixes:

• import math.sqrt more carefully to not worry about overload resolution of sum[].
• Label statistic and p-Value (Pr under null); (It's easy to forget until you see an asterisk mark a "significant" p-value.)
• update dep versions

`fitl/dists.nim` updates

Besides fixing known distro P/F/Q/sampling bugs, add "parabola" & "notch" to round out 0, 1st, 2nd order tail shapes both going down and up at the boundary for eve calibration testing of the new Fraga Alves - Neves estimator.

Seeming stable enough for a minor bump, but don't read too much into version numbers.

Fix sampling for a few key distributions

Fix distribution/cdf/quantile definitions for Triangle and epanechnikov. There are still 5 or 6 broken (cf VC log).

Also, make cross-checking at any x easier & more robust.

Fix non-compilation bug..

..from dumb last minute rename. Even test/polyf.nim did not compile.

Add fitl/svdx.polyFit and test/polyf

Add a not so uncommon special case of a 1-D y(x) weighted polynomial fit and a driver test program (that works fine with test/gen - see comment at top of test/polyf).

Get working with newer nim-devel

That's about it.. also update dependent package versions to latest.

Get rid of most warnings

Get rid of BareExcept, ByteAddress, LineTooLong warnings, unused import, and also sink param makes a copy in hint[Performance] mode.

Re-structure as a nimble package

Re-structure as a (mostly) nimble package and do some minor clean-ups/nim portability things.