The native formatting functions in src/Data/Number/Format.lua passed an
unbound variable `number` to string.format while the actual Number argument
was bound as `n`. At runtime `number` resolved to a nil global, so
toPrecisionNative/toFixedNative/toExponentialNative errored instead of
formatting their input. Reference each function's real argument `n`.

Closes purescript-lua#1

fix: use bound argument n in Format.lua FFI instead of unbound number

…pts/build, run test via bash)

chore: normalize build tooling (purescript-overlay, nix CI, Lua 5.1)

AGENTS.md is the single instruction file the coding agents read natively;
CLAUDE.md is a one-line @AGENTS.md import so Claude Code shares it. Also
removes dead bower.json/package.json/.eslintrc.json. CI is already on the
shared canon; no src/ change.

chore: add AGENTS.md, drop dead JavaScript-era files

Repos moved from Unisay/* to the purescript-lua org. Update the pslua
flake input and lockfile owner (same rev 94c13ce), the packages.dhall
upstream-lua release URL, and the AGENTS.md links, so nothing relies on
the old-owner redirect.

chore: point pslua + package-set links at purescript-lua org

Wire treefmt via treefmt-nix: nixfmt, dhall format, purs-tidy (.tidyrc.json)
and LuaFormatter for the FFI (.lua-format, kept over StyLua because it
preserves the parentheses pslua's parser needs). `nix fmt` formats; the dev
shell installs a content-based pre-commit hook and CI runs
`nix fmt && git diff --exit-code` (content-based, since the in-place
formatters bump mtime and would trip treefmt --fail-on-change). Lua lines
budget 130 cols, matching the raised `luacheck --max-line-length`. The bulk
of the diff is the first format pass.

chore: add treefmt formatting (nix fmt + pre-commit + CI check)

Replace the dev-shell installer that wrote .git/hooks/pre-commit with a
tracked .githooks/pre-commit wired via `git config core.hooksPath .githooks`.

Fixes three issues with the old hook: it gated on `[ -d .git ]` (false in
worktrees/submodules where .git is a file), it clobbered any existing
.git/hooks/pre-commit on every `nix develop`, and `nix fmt … || exit 0`
swallowed real formatter failures. The new hook skips only when `nix` is
absent (CI is the authoritative gate) and otherwise blocks on a `nix fmt`
failure or reformat.

Toolchain consistency: pin the pslua dev flake input to the current
compiler main across the ecosystem. flake.lock only; the package set
consumes sources + FFI, not this dev input, so no re-tag is needed.

string.format("%.<d>d", n) is integer formatting, so toFixed dropped the
fractional part (e.g. toFixed 2 3.14159 -> garbage). Use %f for fixed-point
decimals. Fixes purescript-lua#3.

Verified under Lua 5.1 (3.14159->3.14, 3.7@0->4, 2.5@3->2.500, 10@2->10.00).

Aligns the FFI with the modules' own docstrings (the JS Number semantics):

- toPrecisionNative used %f (d digits after the point); JS toPrecision is
  d SIGNIFICANT digits, so it was wrong for every input and destroyed small
  magnitudes (precision 3 of 0.000123456 gave "0.000"). Now uses %g.
- fromStringImpl wrapped tonumber's nil in Just on a parse failure (a
  Just Number holding nil) and ignored the supplied isFinite predicate. It
  now extracts a leading numeric prefix (matching parseFloat's tolerance of
  a whitespace prefix and trailing junk, e.g. "  1.2 ??" -> Just 1.2),
  guards the parse, and uses isFinite, so "bad"/"" -> Nothing.
- sign encoded the result through the boolean condition, so sign 0 returned
  the boolean `true` instead of the Number 0. Now uses explicit branches.
- max/min folded with Lua's >/< which drop NaN, so e.g. max 1 NaN gave 1;
  JS Math.max/min return NaN if either argument is NaN. Added a NaN guard.
- toExponentialNative emitted C's zero-padded exponent (1.23e+03); JS uses
  the minimal exponent width (1.23e+3). Strips one leading exponent zero.
- toString spelled the specials "inf"/"-inf"/"nan"; now "Infinity"/
  "-Infinity"/"NaN". (Finite values keep Lua's %.14g formatting — the JS
  shortest-round-trip form is out of scope; see #99 / accepted divergences.)

Extends test/regression/format.lua and adds test/regression/number.lua.

Fixes purescript-lua/purescript-lua#92
Fixes purescript-lua/purescript-lua#93
Fixes purescript-lua/purescript-lua#94
Fixes purescript-lua/purescript-lua#95
Fixes purescript-lua/purescript-lua#96
Fixes purescript-lua/purescript-lua#97
Fixes purescript-lua/purescript-lua#98

Build against the published purescript-lua package set via
workspace.packageSet.url instead of the old spago.dhall + packages.dhall
(upstream-ps // upstream-lua) merge. The new spago drives spago.yaml/spago.lock
and dropped Dhall.

- spago.dhall/packages.dhall -> spago.yaml (package name lua-numbers; the new
  spago rejects the purescript- prefix). backend.cmd "true" keeps codegen on
  corefn; scripts/build runs pslua over it as before (dropped the now-redundant
  `-u '-g corefn'` flag since the no-op backend already selects corefn).
- flake.nix: spago 0.21.0 -> 1.x; drop the now-unused dhall dev-shell tool.
- treefmt.nix: drop the dead Dhall formatter. AGENTS.md: spago 1.x, src/packages.json.
- CI action versions bumped to current majors (checkout v5, install-nix v31).
- Remove the upstream PureScript test source test/Test/Main.purs (never built:
  old sources globbed src/ only, and it carries JS FFI that doesn't target Lua).
  The Lua regression tests under test/regression stay.

Dead config: nix fmt formats via treefmt.nix, and nothing reads treefmt.toml
(the pre-commit hook and CI both run nix fmt).

Adds a CHANGELOG.md managed with scriv (fragments in changelog.d/), provides
scriv in the flake dev shell, and records the fork's Lua 5.1 FFI release line.
Where the upstream PureScript changelog existed it is preserved below the fork
section. The AGENTS.md Releasing note now points at the scriv flow instead of
"no changelog entry".

Part of the ecosystem-wide changelog adoption
(purescript-lua/purescript-lua#101); canon in package-set ADR 0009.