carve-lang 0.1.9 → 0.1.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 796f44a79683bfd2b7add107652d71a7d38aaf6d4aedf7169acde4fa2401af26
-  data.tar.gz: d7f0ba287fc9c13eab535e7e376de3096f1baedfd1df415b5f870ec7bfca395f
+  metadata.gz: cd1a19b5989abf287804723f52ff9b4f6ab7098be2d7a6f893caffa7e9ca4f26
+  data.tar.gz: 1c503d1890015fd5f129d8464405a6f66a594517026abbf588474fec96ad8e5e
 SHA512:
-  metadata.gz: aec21e3a2aafe3fd84c6debf48af398ea241610d606b954c770a84bd6b5e38b90b0b1d056ecb4c47f8ba9ec976d85498a5d3013ab6885e1b9fe8fdae572d516b
-  data.tar.gz: 6adc9afe363cd5c135a649b6f336cf41a55b6197a4cb8927273fd57fe4e0c484d47bbe159a8469eea11dc49a41771637ca048712bd596a9bd9515e099e0f611a
+  metadata.gz: 80e49b48040288e254718b55f452a0f0d8a3beab8246a7e50f4f8fb3db50d65597453389a0fcbbdb1610a54f99f51c2659da63b7a60482f700c82fe522b92642
+  data.tar.gz: 03e5140ca7a00cc0e71c5ca0db95cfabfd223cb15d77cfb456373f8538948c6ba6b12e882d0f81843d466b369a4b13fc64ae9cdd19e1fd59ed22ef9b2deac857

data/CHANGELOG.md

@@ -7,6 +7,54 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.1.11] - 2026-10-23
+
+### Improvements
+
+- `Carve.lint(source, extensions:)` exposes the engine's linter: findings carry
+  the rule id, message, 1-based line and column, and character offsets. With
+  `root:` and `source_path:` it lints included files and reports each finding's
+  file and `includedBy` chain (#223).
+
+## [0.1.10] - 2026-10-23
+
+### Breaking
+
+- The engine is the published `carve-lang` 0.1.11 crate. A quote-prefixed `+`
+  needs one space after each `>`, so `>+` and `>>+` now render as text, and a
+  continuation marker reads only at column 0 or right after a block opener.
+  Re-check documents that place `+` elsewhere (#215, markup-carve/carve#3421).
+- Includes of a file that is not valid UTF-8 are refused with an
+  `include-non-text` warning instead of being spliced (#215,
+  markup-carve/carve-rs#2885).
+
+### Fixes
+
+- The carve-lang 0.1.11 pin corrects container ownership for quotes, lazy lines,
+  colon divs and footnotes, fences inside nested lists and quotes, and list
+  tightness around captions (#215).
+- Includes see through a byte order mark and keep link definitions and region
+  markers in place. Markdown import keeps link destinations, autolink brackets
+  and pipe escapes as written (#215).
+- `Carve.render_with_includes` with the `markdown`, `plain` and `ansi` targets
+  keeps a link definition where it was written, as `Carve.to_markdown` does,
+  instead of moving it to the end (#221).
+
+### Improvements
+
+- Precompiled gems for `x86_64-linux`, `x86_64-linux-musl`, `aarch64-linux`,
+  `x86_64-darwin`, `arm64-darwin` and `x64-mingw-ucrt` install without a Rust
+  toolchain on Ruby 3.0 to 4.0 (#214).
+- `Carve.to_carve(source, align_tables: true)` pads table columns to display
+  width (#215).
+- `Carve.select_include_region(source, name, mode:)` resolves a named
+  `@region` to a Range of content lines and raises `Carve::IncludeRegionError`
+  for a missing or malformed region (#215).
+- An include with `@code` or `@code:lang` splices a file or region as one code
+  block (#215).
+- Include warnings carry `includedBy`, the chain of directives that reached the
+  file the warning arose in, with root-relative paths, lines and columns (#221).
+
 ## [0.1.9] - 2026-10-18
 
 ### Breaking
@@ -488,7 +536,9 @@ are not listed, because no release ever shipped them.
   Arrays (every AST node type is covered), enabling custom renderers such as
   [carve-hexapdf](https://github.com/markup-carve/carve-hexapdf).
 
-[Unreleased]: https://github.com/markup-carve/carve-rb/compare/v0.1.9...HEAD
+[Unreleased]: https://github.com/markup-carve/carve-rb/compare/v0.1.11...HEAD
+[0.1.11]: https://github.com/markup-carve/carve-rb/compare/v0.1.10...v0.1.11
+[0.1.10]: https://github.com/markup-carve/carve-rb/compare/v0.1.9...v0.1.10
 [0.1.9]: https://github.com/markup-carve/carve-rb/compare/v0.1.8...v0.1.9
 [0.1.8]: https://github.com/markup-carve/carve-rb/compare/v0.1.7...v0.1.8
 [0.1.7]: https://github.com/markup-carve/carve-rb/compare/v0.1.6...v0.1.7

data/README.md

@@ -16,7 +16,48 @@ bundle install
 ```
 
 The distribution name is `carve-lang`, while the require path is `carve`.
-Building from source requires Rust 1.75 or newer and Ruby development headers.
+
+Releases after 0.1.9 ship precompiled gems for `x86_64-linux`,
+`x86_64-linux-musl`, `aarch64-linux`, `x86_64-darwin`, `arm64-darwin` and
+`x64-mingw-ucrt`, with a binary for each Ruby from 3.0 to 4.0 (from 3.1 on
+Windows, where Ruby 3.0 has no UCRT build). On those platforms `gem install`
+needs no Rust toolchain or libclang. Everywhere else, including newer Rubies,
+RubyGems installs the source gem and builds it as described below.
+
+## Installing from source
+
+Version 0.1.9 and earlier ship source only, and so does every release on a
+platform or Ruby without a precompiled gem. There `gem install carve-lang`
+compiles a Rust extension on your machine, so the install needs:
+
+- A Rust toolchain with `cargo`, Rust 1.77 or newer. [rustup](https://rustup.rs)
+  is the usual way to get one.
+- Ruby development headers (`ruby-dev` on Debian and Ubuntu).
+- libclang and clang. The build runs `bindgen` through `rb-sys` to read Ruby's
+  C headers. On Debian and Ubuntu:
+
+```bash
+sudo apt install libclang-dev clang
+```
+
+If only libclang is installed, without the `clang` package, the build stops
+with this error:
+
+```text
+/usr/include/ruby-3.2.0/ruby/ruby.h:23:10: fatal error: 'stdarg.h' file not found
+thread 'main' panicked at .../rb-sys-0.9.128/build/main.rs:52:6:
+generate bindings: ClangDiagnostic("/usr/include/ruby-3.2.0/ruby/ruby.h:23:10: fatal error: 'stdarg.h' file not found\n")
+error: failed to run custom build command for `rb-sys v0.9.128`
+```
+
+libclang cannot find its own builtin headers. Install `clang`, or point
+bindgen at GCC's copy of those headers:
+
+```bash
+BINDGEN_EXTRA_CLANG_ARGS="-I$(gcc -print-file-name=include)" gem install carve-lang
+```
+
+With Bundler, export the variable before running `bundle install`.
 
 ## Render Carve
 
@@ -28,6 +69,7 @@ Carve.to_markdown(source)
 Carve.to_plain_text(source)
 Carve.to_ansi(source)
 Carve.to_carve(source)
+Carve.to_carve(source, align_tables: true) # pad table columns
 ```
 
 Options configure extensions, safe rendering, profiles, symbols, section
@@ -91,11 +133,28 @@ result = Carve.to_html_with_includes(
 )
 ```
 
-The result contains the value, warnings, and root-relative dependencies.
+The result contains the value, warnings, and root-relative dependencies. A
+warning from an included file names the chain of directives that reached it in
+`:includedBy`.
 Other render methods leave directives literal. See the
 [include reference](docs/reference.md#file-includes) for budgets, AST expansion,
 targets, and containment behavior.
 
+## Lint
+
+`Carve.lint(source)` reports constructs that render but lose what the author
+wrote, such as an unattached attribute or a link to a missing id:
+
+```ruby
+Carve.lint("é [x](#nope)\n").map { |f| [f[:rule], f[:line], f[:column]] }
+# => [["broken-fragment-link", 1, 3]]
+```
+
+Columns and the `:start`/`:end` offsets count characters, not bytes. Pass
+`root:` and `source_path:` to lint included files too; each such finding names
+its file and the `:includedBy` chain. See the
+[lint reference](docs/reference.md#linting).
+
 ## Security
 
 Use `safe: true` and a restrictive profile for untrusted documents. Custom

data/ext/carve/Cargo.lock

@@ -37,9 +37,9 @@ checksum = "b4388bee8683e3d04af747c73422af53102d2bd24d9eadb6cbc100baef4b43f8"
 
 [[package]]
 name = "carve-lang"
-version = "0.1.10"
+version = "0.1.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4a9b7a2a3b758f76ce1888ad1371bd933c5f2ccded01305c03f1550641a45c1d"
+checksum = "753b90fbe80fab43887d5d559996c153347d9631452bbcb8d0283aad6dfa7ef8"
 dependencies = [
  "html5ever",
  "markup5ever_rcdom",
@@ -54,7 +54,7 @@ dependencies = [
 
 [[package]]
 name = "carve-rb"
-version = "0.1.9"
+version = "0.1.11"
 dependencies = [
  "carve-lang",
  "magnus",

data/ext/carve/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "carve-rb"
-version = "0.1.9"
+version = "0.1.11"
 edition = "2021"
 rust-version = "1.75"
 publish = false
@@ -22,4 +22,4 @@ serde_json = "1"
 # The Carve engine. Imported under the Rust crate alias `carve_rs` so the
 # binding's own crate name (`carve-rb`) does not collide with the engine's
 # package name (`carve`) in the Cargo lockfile.
-carve_rs = { package = "carve-lang", version = "=0.1.10" }
+carve_rs = { package = "carve-lang", version = "=0.1.11" }

data/ext/carve/src/lib.rs

@@ -226,8 +226,44 @@ fn to_ansi(source: String) -> String {
     carve_rs::to_ansi(&source)
 }
 
-fn to_carve(source: String) -> String {
-    carve_rs::to_carve(&source)
+/// Canonical Carve source; a writer refusal returns the source as authored.
+fn to_carve(source: String, align_tables: bool) -> String {
+    let options = Options::default().with_align_tables(align_tables);
+    carve_rs::try_to_carve_with_options(&source, &options).unwrap_or(source)
+}
+
+/// Resolve `@region:name` with the engine's marker rules (spec I1b).
+///
+/// Returns `[start, end]` (1-based, inclusive content lines) or, when the
+/// region is invalid, the warning rule id as a String; `lib/carve.rb` turns
+/// that into `Carve::IncludeRegionError`.
+fn select_include_region(
+    ruby: &Ruby,
+    source: String,
+    name: String,
+    mode: String,
+) -> Result<Value, Error> {
+    let mode = match mode.as_str() {
+        "carve" => carve_rs::IncludeRegionMode::Carve,
+        "code" => carve_rs::IncludeRegionMode::Code,
+        other => {
+            return Err(Error::new(
+                ruby.exception_arg_error(),
+                format!("mode must be :carve or :code, got {other:?}"),
+            ))
+        }
+    };
+    Ok(
+        match carve_rs::select_include_region(&source, &name, mode) {
+            Ok(region) => ruby
+                .ary_new_from_values(&[
+                    ruby.integer_from_u64(region.start as u64).as_value(),
+                    ruby.integer_from_u64(region.end as u64).as_value(),
+                ])
+                .as_value(),
+            Err(e) => ruby.str_new(e.rule()).as_value(),
+        },
+    )
 }
 
 /// Which renderer runs over the expanded document.
@@ -261,28 +297,16 @@ fn parse_include_target(ruby: &Ruby, name: &str) -> Result<IncludeTarget, Error>
     }
 }
 
-#[allow(clippy::too_many_arguments)]
-fn render_with_includes_json(
+/// Canonicalize an include root and the document's own path, and refuse a
+/// document outside the root. Both must be absolute: a relative one would root
+/// containment at the process working directory.
+fn contained_include_paths(
     ruby: &Ruby,
-    source: String,
-    root: String,
-    source_path: String,
-    target: String,
-    names: RArray,
-    mode: String,
-    renderers: RHash,
-    symbols: RHash,
-    safe: bool,
-    profile: Option<String>,
-    sections: bool,
-    max_depth: Option<usize>,
-    max_bytes: Option<usize>,
-    max_resolver_calls: Option<usize>,
-    max_warnings: Option<usize>,
-) -> Result<String, Error> {
-    let which = parse_include_target(ruby, &target)?;
-    let root_path = Path::new(&root);
-    let source_path = Path::new(&source_path);
+    root: &str,
+    source_path: &str,
+) -> Result<(PathBuf, PathBuf), Error> {
+    let root_path = Path::new(root);
+    let source_path = Path::new(source_path);
     if !root_path.is_absolute() || !source_path.is_absolute() {
         return Err(Error::new(
             ruby.exception_arg_error(),
@@ -307,6 +331,42 @@ fn render_with_includes_json(
             "include source_path must be a file inside the include root".to_string(),
         ));
     }
+    Ok((root_real, source_real))
+}
+
+/// A path as a reader sees it: relative to the include root.
+fn root_relative(root_real: &Path, value: &str) -> String {
+    let path = PathBuf::from(value);
+    if path.is_absolute() {
+        path.strip_prefix(root_real)
+            .map(|relative| relative.to_string_lossy().replace('\\', "/"))
+            .unwrap_or_else(|_| "[outside-root]".to_string())
+    } else {
+        value.to_string()
+    }
+}
+
+#[allow(clippy::too_many_arguments)]
+fn render_with_includes_json(
+    ruby: &Ruby,
+    source: String,
+    root: String,
+    source_path: String,
+    target: String,
+    names: RArray,
+    mode: String,
+    renderers: RHash,
+    symbols: RHash,
+    safe: bool,
+    profile: Option<String>,
+    sections: bool,
+    max_depth: Option<usize>,
+    max_bytes: Option<usize>,
+    max_resolver_calls: Option<usize>,
+    max_warnings: Option<usize>,
+) -> Result<String, Error> {
+    let which = parse_include_target(ruby, &target)?;
+    let (root_real, source_real) = contained_include_paths(ruby, &root, &source_path)?;
 
     let resolver = carve_rs::FileSystemResolver::new(&root_real).map_err(|_| {
         Error::new(
@@ -344,6 +404,12 @@ fn render_with_includes_json(
     if !sections {
         render_options = render_options.with_sections(false);
     }
+    // Positions let the engine record the directive chain behind a warning
+    // (`includedBy`). The AST target keeps them off: spec I4 leaves position
+    // remapping out of scope, so a child's spans would point into the wrong file.
+    if which != IncludeTarget::Ast {
+        render_options = render_options.with_positions(true);
+    }
     let mut include_options = carve_rs::IncludeOptions::new()
         .with_resolver(&resolver)
         .with_source_path(source_real.to_string_lossy());
@@ -401,16 +467,7 @@ fn render_with_includes_json(
     } else {
         serde_json::Value::String(rendered)
     };
-    let safe_path = |value: &str| {
-        let path = PathBuf::from(value);
-        if path.is_absolute() {
-            path.strip_prefix(&root_real)
-                .map(|relative| relative.to_string_lossy().replace('\\', "/"))
-                .unwrap_or_else(|_| "[outside-root]".to_string())
-        } else {
-            value.to_string()
-        }
-    };
+    let safe_path = |value: &str| root_relative(&root_real, value);
     let warnings = prepared
         .warnings
         .iter()
@@ -419,6 +476,17 @@ fn render_with_includes_json(
                 "rule": warning.rule,
                 "message": warning.message,
                 "file": warning.file.as_deref().map(&safe_path),
+                "includedBy": warning
+                    .included_by
+                    .iter()
+                    .map(|site| {
+                        serde_json::json!({
+                            "file": safe_path(&site.file),
+                            "line": site.line,
+                            "column": site.column,
+                        })
+                    })
+                    .collect::<Vec<_>>(),
             })
         })
         .collect::<Vec<_>>();
@@ -442,6 +510,122 @@ fn render_with_includes_json(
     .to_string())
 }
 
+/// Byte offset -> codepoint offset over `source`.
+///
+/// `LintWarning` reports `start`/`end` in bytes so a Rust caller can slice its
+/// `&str`; a Ruby String indexes by character, so a byte offset would mis-slice
+/// every document with a non-ASCII character before the finding. A byte inside a
+/// character maps to that character, and the final slot is the total.
+fn codepoint_table(source: &str) -> Vec<usize> {
+    let mut table = Vec::with_capacity(source.len() + 1);
+    let mut count = 0usize;
+    for ch in source.chars() {
+        for _ in 0..ch.len_utf8() {
+            table.push(count);
+        }
+        count += 1;
+    }
+    table.push(count);
+    table
+}
+
+#[allow(clippy::too_many_arguments)]
+fn lint_json(
+    ruby: &Ruby,
+    source: String,
+    names: RArray,
+    root: Option<String>,
+    source_path: Option<String>,
+    max_depth: Option<usize>,
+    max_bytes: Option<usize>,
+    max_resolver_calls: Option<usize>,
+    max_warnings: Option<usize>,
+) -> Result<String, Error> {
+    let boxed = boxed_extensions(ruby, names)?;
+    let mut options = Options::new();
+    for extension in &boxed {
+        options = options.with_extension(extension.as_ref());
+    }
+    // The document's own findings come first in `lint_carve_with_options`, and
+    // they are the only ones whose offsets index `source`. Linting it alone
+    // first tells the two apart without guessing from a finding's fields.
+    let own = carve_rs::lint_carve_with_options(&source, &options);
+    let own_count = own.len();
+    let (warnings, root_real) = match (root, source_path) {
+        (None, None) => (own, None),
+        (Some(root), Some(source_path)) => {
+            let (root_real, source_real) = contained_include_paths(ruby, &root, &source_path)?;
+            let resolver = carve_rs::FileSystemResolver::new(&root_real).map_err(|_| {
+                Error::new(
+                    ruby.exception_arg_error(),
+                    "include root is not usable".to_string(),
+                )
+            })?;
+            let mut include_options = carve_rs::IncludeOptions::new()
+                .with_resolver(&resolver)
+                .with_source_path(source_real.to_string_lossy());
+            if let Some(value) = max_depth {
+                include_options = include_options.with_max_depth(value);
+            }
+            if let Some(value) = max_bytes {
+                include_options = include_options.with_max_bytes(value);
+            }
+            if let Some(value) = max_resolver_calls {
+                include_options = include_options.with_max_resolver_calls(value);
+            }
+            if let Some(value) = max_warnings {
+                include_options = include_options.with_max_warnings(value);
+            }
+            let options = options.with_lint_includes(include_options);
+            (
+                carve_rs::lint_carve_with_options(&source, &options),
+                Some(root_real),
+            )
+        }
+        _ => {
+            return Err(Error::new(
+                ruby.exception_arg_error(),
+                "lint needs both root: and source_path: to follow includes".to_string(),
+            ))
+        }
+    };
+    let table = codepoint_table(&source);
+    let total = table.len() - 1;
+    let to_codepoint = |byte: usize| table.get(byte).copied().unwrap_or(total);
+    let path = |value: &str| match &root_real {
+        Some(root_real) => root_relative(root_real, value),
+        None => value.to_string(),
+    };
+    let findings = warnings
+        .iter()
+        .enumerate()
+        .map(|(index, warning)| {
+            let own = index < own_count;
+            serde_json::json!({
+                "rule": warning.rule,
+                "message": warning.message,
+                "line": warning.line,
+                "column": warning.column,
+                "start": own.then(|| to_codepoint(warning.start)),
+                "end": own.then(|| to_codepoint(warning.end)),
+                "file": warning.file.as_deref().map(&path),
+                "includedBy": warning
+                    .included_by
+                    .iter()
+                    .map(|site| {
+                        serde_json::json!({
+                            "file": path(&site.file),
+                            "line": site.line,
+                            "column": site.column,
+                        })
+                    })
+                    .collect::<Vec<_>>(),
+            })
+        })
+        .collect::<Vec<_>>();
+    Ok(serde_json::Value::Array(findings).to_string())
+}
+
 fn from_html_json(ruby: &Ruby, source: String, mode: String) -> Result<String, Error> {
     let mode = carve_rs::HtmlImportMode::from_name(&mode).ok_or_else(|| {
         Error::new(
@@ -897,7 +1081,11 @@ guarded!(g_to_html(source: Source) -> String => to_html);
 guarded!(g_to_markdown(source: Source) -> String => to_markdown);
 guarded!(g_to_plain_text(source: Source) -> String => to_plain_text);
 guarded!(g_to_ansi(source: Source) -> String => to_ansi);
-guarded!(g_to_carve(source: Source) -> String => to_carve);
+guarded!(g_to_carve(source: Source, align_tables: bool) -> String => to_carve);
+guarded!(
+    g_select_include_region(ruby: &Ruby, source: Source, name: String, mode: String) -> Value
+        =>? select_include_region
+);
 guarded!(g_from_markdown_json(ruby: &Ruby, source: Source) -> String =>? from_markdown_json);
 guarded!(g_to_ast_json(source: Source) -> String => to_ast_json);
 guarded!(g_extension_names() -> Vec<String> => extension_names);
@@ -967,6 +1155,20 @@ guarded!(
     ) -> String =>? render_with_includes_json
 );
 
+guarded!(
+    g_lint_json(
+        ruby: &Ruby,
+        source: Source,
+        names: RArray,
+        root: Option<String>,
+        source_path: Option<String>,
+        max_depth: Option<usize>,
+        max_bytes: Option<usize>,
+        max_resolver_calls: Option<usize>,
+        max_warnings: Option<usize>
+    ) -> String =>? lint_json
+);
+
 /// Entry point invoked by Ruby when the extension is loaded.
 ///
 /// `name = "carve"` makes the macro emit the `Init_carve` symbol that matches
@@ -990,7 +1192,11 @@ fn init(ruby: &Ruby) -> Result<(), Error> {
     module.define_singleton_method("to_markdown", function!(g_to_markdown, 1))?;
     module.define_singleton_method("to_plain_text", function!(g_to_plain_text, 1))?;
     module.define_singleton_method("to_ansi", function!(g_to_ansi, 1))?;
-    module.define_singleton_method("to_carve", function!(g_to_carve, 1))?;
+    module.define_singleton_method("_to_carve", function!(g_to_carve, 2))?;
+    module.define_singleton_method(
+        "_select_include_region",
+        function!(g_select_include_region, 3),
+    )?;
     module.define_singleton_method(
         "_render_with_includes_json",
         function!(g_render_with_includes_json, 15),
@@ -1012,5 +1218,6 @@ fn init(ruby: &Ruby) -> Result<(), Error> {
     module.define_singleton_method("_read_stamp", function!(g_read_stamp, 1))?;
     module.define_singleton_method("_stamp_needs_review", function!(g_stamp_needs_review, 2))?;
     module.define_singleton_method("_panic_probe", function!(g_panic_probe, 0))?;
+    module.define_singleton_method("_lint_json", function!(g_lint_json, 8))?;
     Ok(())
 }

data/lib/carve/version.rb

@@ -8,5 +8,5 @@ module Carve
   # compares it against ext/carve/Cargo.toml and against the newest cut
   # CHANGELOG section on every run, and .github/workflows/release.yml refuses to
   # publish a gem whose version is not the tag being released.
-  VERSION = "0.1.9"
+  VERSION = "0.1.11"
 end

data/lib/carve.rb

@@ -8,11 +8,30 @@
 require "json"
 require_relative "carve/version"
 
-# Load the compiled native extension. Built by `rake compile` to
-# lib/carve/carve.so (or .bundle on macOS).
-require_relative "carve/carve"
+# Load the compiled native extension. A precompiled platform gem ships one per
+# Ruby ABI under lib/carve/<major.minor>/; a source install and `rake compile`
+# build lib/carve/carve.so (or .bundle on macOS).
+# Chosen by directory rather than by rescuing LoadError, so a binary that is
+# present but cannot load (an older glibc, say) reports its own error.
+abi = RUBY_VERSION[/\A\d+\.\d+/]
+if File.directory?(File.join(__dir__, "carve", abi))
+  require_relative "carve/#{abi}/carve"
+else
+  require_relative "carve/carve"
+end
 
 module Carve
+  # Raised by .select_include_region when the region is missing or a marker is
+  # malformed, unpaired or duplicated. +rule+ is the engine's warning rule id.
+  class IncludeRegionError < ArgumentError
+    attr_reader :rule
+
+    def initialize(rule, name)
+      @rule = rule
+      super("#{rule}: region #{name.inspect} is missing or its markers are malformed")
+    end
+  end
+
   # Extension names the native binding understands.
   #
   # Taken from the engine itself rather than typed out here. This was a literal
@@ -55,7 +74,8 @@ module Carve
     #
     # Returns +{value:, warnings:, dependencies:, suppressedWarnings:}+. Paths
     # in +warnings+ and +dependencies+ are relative to the root, so a report can
-    # be shown to a reader as it stands.
+    # be shown to a reader as it stands. Each warning's +includedBy+ lists the
+    # directives that reached its file, top-level first, as +{file:, line:, column:}+.
     #
     # Every render option +to_html+ takes is accepted and reaches the included
     # children too: an extension, a symbol map or a profile means the same thing
@@ -130,6 +150,77 @@ module Carve
       )
     end
 
+    # Report constructs that parse and render, but not the way the author
+    # meant: an attribute that attaches to nothing, an unclosed fence, a
+    # fragment link with no target.
+    #
+    #   Carve.lint("Text.\n\n{#gone}\n")
+    #   # => [{rule: "unattached-block-attribute", message: "...", line: 3,
+    #   #      column: 1, start: 7, end: 14, file: nil, includedBy: []}]
+    #
+    # Returns an Array of Hashes, empty for a clean document:
+    #
+    # * +:rule+ - the engine's stable rule id, the same in every Carve engine.
+    # * +:message+ - what degrades, in prose.
+    # * +:line+, +:column+ - 1-based; the column counts characters, not bytes.
+    # * +:start+, +:end+ - 0-based character offsets into +source+, end
+    #   exclusive, so <tt>source[f[:start]...f[:end]]</tt> is the offending text.
+    #   +nil+ on a finding the include pass adds (below), whose offsets would
+    #   index another text.
+    # * +:file+, +:includedBy+ - see below; +nil+ and +[]+ without includes.
+    #
+    # The engine reports no severity: every finding is a warning.
+    #
+    # Pass the same +extensions:+ you render with, since they change which
+    # constructs are elements and which rules apply.
+    #
+    # ==== Following includes
+    #
+    # With +root:+ and +source_path:+ (absolute, as for .to_html_with_includes)
+    # the pass also lints every included Carve file and reports include
+    # directives that stay literal (a missing target, a cycle, a budget). Such a
+    # finding names its +:file+ relative to the root and its +:includedBy+ chain,
+    # top-level first, as +{file:, line:, column:}+. The budgets
+    # (+max_depth:+ and the rest) mean what they mean for rendering.
+    def lint(source, extensions: nil, root: nil, source_path: nil, max_depth: nil,
+             max_bytes: nil, max_resolver_calls: nil, max_warnings: nil)
+      JSON.parse(
+        _lint_json(
+          source.to_s,
+          Array(extensions).map(&:to_s),
+          root&.to_s,
+          source_path&.to_s,
+          max_depth,
+          max_bytes,
+          max_resolver_calls,
+          max_warnings,
+        ),
+        symbolize_names: true,
+      )
+    end
+
+    # Format +source+ into canonical Carve. A writer refusal returns the source
+    # as authored. +align_tables:+ pads table cells into aligned columns.
+    def to_carve(source, align_tables: false)
+      _to_carve(source, align_tables ? true : false)
+    end
+
+    # Resolve <tt>@region:name</tt> in a file's text with the engine's marker
+    # rules, so an editor jumping to a region agrees with the render.
+    #
+    #   Carve.select_include_region(text, "setup", mode: :carve) # => 3..4
+    #
+    # Returns the 1-based, inclusive Range of content lines; an empty region is
+    # an empty Range such as +2..1+. +mode: :carve+ accepts only
+    # <tt>%% #region</tt> markers, +mode: :code+ also the source-comment forms
+    # of an +@code+ include. Raises IncludeRegionError otherwise.
+    def select_include_region(source, name, mode:)
+      result = _select_include_region(source, name.to_s, mode.to_s)
+      raise IncludeRegionError.new(result, name.to_s) if result.is_a?(String)
+
+      result[0]..result[1]
+    end
+
     # Import HTML or Markdown into canonical Carve. Both methods return the
     # shared migration shape `{ value:, report: }`.
     def from_html(source, mode: :safe)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: carve-lang
 version: !ruby/object:Gem::Version
-  version: 0.1.9
+  version: 0.1.11
 platform: ruby
 authors:
 - markup-carve
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-10-18 00:00:00.000000000 Z
+date: 2026-10-23 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys