feat(lint): add 10 Fennel, Janet and CL declaration rules

Cargo.toml

@@ -142,6 +142,8 @@ paredit-feature-lint-elisp-idiom = { path = "packages/feature/lint-elisp-idiom"
 paredit-feature-lint-pathname-io = { path = "packages/feature/lint-pathname-io" }
 paredit-feature-lint-clojure-idiom = { path = "packages/feature/lint-clojure-idiom" }
 paredit-feature-lint-scheme-idiom = { path = "packages/feature/lint-scheme-idiom" }
+paredit-feature-lint-fennel-janet-idiom = { path = "packages/feature/lint-fennel-janet-idiom" }
+paredit-feature-lint-type-declaration = { path = "packages/feature/lint-type-declaration" }
 paredit-feature-emacs-lisp = { path = "packages/feature/emacs-lisp" }
 paredit-feature-conditional-conversion = { path = "packages/feature/conditional-conversion" }
 paredit-feature-external-check = { path = "packages/feature/external-check" }

docs/src/reference/api.md

@@ -421,7 +421,7 @@ adds each finding's full field set as indented lines under its row.
 
 ### Choosing and tuning lint rules
 
-With 303 rules, `inspect lint` needs more than an on/off switch per rule. The
+With 313 rules, `inspect lint` needs more than an on/off switch per rule. The
 flags below are about the rule *set* rather than about any one rule, and all of
 them work with `--list-rules` as well as with a scan — so a run can be
 inspected before it is made.
@@ -446,7 +446,7 @@ key for baselines and suppression tooling.
 
 ### Rules a project writes for itself
 
-The 303 shipped rules are the ones everybody gets. A rule like "in *this*
+The 313 shipped rules are the ones everybody gets. A rule like "in *this*
 codebase, `defentity` must always be given a `:table`" is the majority of what
 a mature project wants and none of what a linter can ship, so a project writes
 those itself, in Lisp, in `.paredit/rules/*.lisp`:

docs/src/reference/architecture.md

@@ -1,6 +1,6 @@
 # Architecture
 
-`paredit-cli` is a Cargo workspace: a thin composition root plus 62 packages
+`paredit-cli` is a Cargo workspace: a thin composition root plus 64 packages
 under `packages/core/` and `packages/feature/`. Knowing which package owns a
 thing is the fastest way to know where a change belongs.
 
@@ -17,7 +17,7 @@ core/syntax ──▶ core/semantics ──▶ core/edit ──▶ core/cli
      └──▶ core/workspace          core/lint-engine ──┘
                     │
                     ▼
-              feature/*  (53 packages, mostly independent of each other)
+              feature/*  (55 packages, mostly independent of each other)
                     │
                     ▼
               paredit-cli  (command tree, dispatch, REGISTRY)
@@ -65,10 +65,10 @@ src/
 A contract test walks `src/` and refuses anything else.
 
 The lint `REGISTRY` is the canonical example of what *must* live here. It names
-all 303 rules, and every rule depends on the engine; putting the registry in
+all 313 rules, and every rule depends on the engine; putting the registry in
 either would be a cycle. So the engine takes a `RuleCatalog` as an argument and
 never learns which rules exist, the rules never learn the registry does, and
-the registry sits in the root reaching twenty-eight feature packages for their
+the registry sits in the root reaching thirty feature packages for their
 `META` and `RULE`. That is the criterion: **a module that enumerates or
 aggregates several features** belongs in neither core nor any one feature.
 
@@ -127,7 +127,7 @@ semantic enum (`ReportLimit::{Complete, Limited(NonZeroUsize)}`,
 Derive redundant presentation values (booleans, counts) at the serialization
 boundary instead of storing them.
 
-## Lint rules: one trait, one registry line, twenty-eight packages
+## Lint rules: one trait, one registry line, thirty packages
 
 The lint suite is the clearest example of the split's shape, and the most
 frequently extended part of the tree.
@@ -141,8 +141,8 @@ frequently extended part of the tree.
 | `policy` | Dialect scope, rule selection and gate decisions: logic that needs no tree. |
 | `engine` | The single pass, which walks the document once and dispatches each node to every rule whose `head_filter` matches. |
 
-298 of the 303 shipped rules live in twenty-seven themed packages, split seven
-ways. A twenty-eighth, `feature/lint-custom`, holds no rules at all: it is the
+308 of the 313 shipped rules live in twenty-nine themed packages, split seven
+ways. A thirtieth, `feature/lint-custom`, holds no rules at all: it is the
 pattern language and the second pass that run the rules a *project* writes for
 itself.
 
@@ -321,7 +321,7 @@ specifies the two cases R7RS 6.4 leaves open — fixnums compare `eq?` by
 guarantee and characters have been normatively `eq?` since 9.0.0.10 — so every
 finding there would complain about code the language promises will work.
 
-**`REGISTRY` is in neither.** It names all 303 rules, and every rule depends on
+**`REGISTRY` is in neither.** It names all 313 rules, and every rule depends on
 the engine, so putting it in the engine or in a rule package would be a cycle.
 It sits in the root crate, and the engine receives a `RuleCatalog` as an
 argument — which is why the engine can be a package at all.

docs/src/reference/configuration.md

@@ -1,6 +1,6 @@
 # Configuration
 
-With 303 lint rules and 460 commands, passing every knob as a flag
+With 313 lint rules and 460 commands, passing every knob as a flag
 stopped scaling. `paredit.toml` is the answer: a small, strictly validated file
 that sets the defaults a repository wants, so a command line carries only what
 is unusual about *this* invocation.

packages/feature/lint-fennel-janet-idiom/Cargo.toml

@@ -0,0 +1,19 @@
+[package]
+name = "paredit-feature-lint-fennel-janet-idiom"
+description = "Lint rules for Fennel and Janet idiom, the two dialects the catalogue had no dedicated rules for"
+readme = "README.md"
+publish = false
+version.workspace = true
+edition.workspace = true
+rust-version.workspace = true
+license.workspace = true
+repository.workspace = true
+
+[dependencies]
+paredit-core-syntax = { path = "../../core/syntax" }
+paredit-core-lint-engine = { path = "../../core/lint-engine" }
+
+# Mandatory: without it this package silently opts out of the workspace lint
+# table, including `unsafe_code = "deny"`, with no error at all.
+[lints]
+workspace = true

packages/feature/lint-fennel-janet-idiom/README.md

@@ -0,0 +1,51 @@
+# paredit-feature-lint-fennel-janet-idiom
+
+Lint rules whose subject is Fennel or Janet specifically.
+
+Both dialects are first-class in the parser, and until this package the
+catalogue had seven rules in scope for either of them — six of which model
+several dialects at once and none of which encodes a fact about Fennel or Janet
+that is not also true elsewhere. Every rule here is keyed on something the
+language's own reference, compiler, or shipped linter states.
+
+| rule | dialects | keyed on | primary source |
+| --- | --- | --- | --- |
+| `var-never-set` | Fennel, Janet | `var`, `var-` | Fennel `src/linter.fnl` `check-unused`, `"declared as var but never set"` |
+| `fennel-deprecated-form` | Fennel | `global`, `require-macros`, `pick-args` | `reference.md`, "Deprecated Forms" |
+| `fennel-each-over-non-iterator` | Fennel | `each` | `specials.fnl` `SPECIALS.each`, which emits Lua's generic `for … in` |
+| `janet-empty-loop-body` | Janet | `loop`, `seq`, `catseq` | `boot.janet` `check-empty-body`, `maclintf :normal "empty loop body"` |
+| `janet-mutating-immutable-literal` | Janet | `put`, `array/*`, `buffer/*` | `src/core/value.c` `janet_put`, which panics on a struct |
+
+## Third-party audit
+
+Run over code nobody here wrote: 288 `.fnl` files (`fennel-lang/fennel`,
+`Olical/conjure`, `rktjmp/hotpot.nvim`, `udayvir-singh/tangerine.nvim`,
+`min-love2d-fennel`) and 241 `.janet` files (`janet-lang/janet`, `spork`,
+`jpm`, `circlet`, `andrewchambers/janet-sh`).
+
+| rule | candidates | findings | adjudication |
+| --- | --- | --- | --- |
+| `var-never-set` (Fennel) | 210 | 8 | all true; 5 more were false positives from a project-local macro expanding to `set`, now suppressed |
+| `var-never-set` (Janet) | 464 | 31 | all true; 5 more were the same macro false positive, now suppressed |
+| `fennel-deprecated-form` | 29 | 28 | all true; the 29th is a malformed `(global)` the arity guard declines |
+| `fennel-each-over-non-iterator` | 194 | 1 | true — and it is `fennel-lang/fennel`'s own assertion that this shape raises |
+| `janet-empty-loop-body` | 174 | 0 | 2 findings before the `:iterate` narrowing, both the deliberate drain idiom |
+| `janet-mutating-immutable-literal` | 695 | 0 | unproven: a real denominator, no instance in this corpus |
+
+## Known parser limitation for Janet
+
+Janet's long strings are delimited by a run of backticks of any length
+(`src/core/parse.c`, `longstring`), and this repository's reader does not
+implement them: a backtick is neither a delimiter nor whitespace for
+`Dialect::Janet`, so it is absorbed into an atom and the string's contents are
+read as code. Over 241 files from `janet-lang/janet`, `spork`, `jpm`, `circlet`
+and `janet-sh`, 9 fail to parse outright and 41 more parse into a tree
+containing nodes that lie inside a long string's body. Docstrings written with
+```` ``` ```` are the dominant cause, and `src/boot/boot.janet` — Janet's own
+core library — is one of the nine.
+
+Every rule here is therefore blind on roughly 4% of real Janet files and can be
+handed phantom forms on another 17%. The rules' quote guard does not help:
+these nodes are not quoted, they are prose. Fixing the reader is out of this
+package's scope; the measurement is recorded here so the limitation is not
+rediscovered.