RubyGems - html-to-markdown - Versions diffs - 2.25.1 → 2.26.2 - Mend

html-to-markdown 2.25.1 → 2.26.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (2075) hide show

data/rust-vendor/html-to-markdown-rs/tests/visitor_integration_test.rs CHANGED Viewed

@@ -921,3 +921,61 @@ fn test_convert_with_all_features_and_visitor() {
     // Verify markdown was produced
     assert!(!result.markdown.is_empty(), "Should produce markdown output");
 }
+/// Regression test: image visitor returning Custom with metadata extraction used to panic
+/// with an out-of-bounds slice.
+///
+/// When metadata extraction prepends a YAML frontmatter block to `output`, every element's
+/// saved `element_output_start` is offset by the frontmatter length.  If a child visitor
+/// then returns Custom and truncates the buffer, the parent's saved offset can point
+/// past `output.len()`.
+#[test]
+fn test_image_visitor_with_metadata_does_not_panic() {
+    #[derive(Debug)]
+    struct ImageVisitor;
+    impl HtmlVisitor for ImageVisitor {
+        fn visit_image(&mut self, _ctx: &NodeContext, _src: &str, _alt: &str, _title: Option<&str>) -> VisitResult {
+            VisitResult::Custom("![img](rewritten.png)".to_string())
+        }
+    }
+    let html = r#"<html><head><meta name="description" content="x"></head><body><p><img src="a.png" alt="a"></p></body></html>"#;
+    let options = ConversionOptions {
+        extract_metadata: true,
+        ..Default::default()
+    };
+    let result = convert_with_visitor(html, Some(options), Some(Rc::new(RefCell::new(ImageVisitor))));
+    assert!(result.is_ok(), "conversion panicked or errored: {:?}", result.err());
+}
+/// Regression test: visit_element_end returning Custom/Skip with metadata extraction used
+/// to produce stale parent offsets and either panic or silently drop subsequent content.
+#[test]
+fn test_element_end_replacement_with_metadata_preserves_subsequent_content() {
+    #[derive(Debug)]
+    struct FigureReplacingVisitor;
+    impl HtmlVisitor for FigureReplacingVisitor {
+        fn visit_element_end(&mut self, ctx: &NodeContext, _content: &str) -> VisitResult {
+            if ctx.tag_name == "figure" {
+                return VisitResult::Custom("[figure]".to_string());
+            }
+            VisitResult::Continue
+        }
+    }
+    let html = r#"<html><head><meta name="description" content="x"></head><body><figure><img src="a.png"></figure><p>after</p></body></html>"#;
+    let options = ConversionOptions {
+        extract_metadata: true,
+        ..Default::default()
+    };
+    let result = convert_with_visitor(html, Some(options), Some(Rc::new(RefCell::new(FigureReplacingVisitor))));
+    assert!(result.is_ok(), "conversion panicked or errored: {:?}", result.err());
+    assert!(
+        result.unwrap().contains("after"),
+        "content after replaced element should not be lost"
+    );
+}

data/rust-vendor/html5ever/.cargo-checksum.json CHANGED Viewed

	@@ -1 +1 @@
1	- {"files":{".cargo_vcs_info.json":"~~ccf5bcae2f8886fef19836ac305a9181987d0fbc3a361f8acc0ea898d8b635f9~~","Cargo.lock":"~~edaa4d682377a9786d9ffaf70575ccc5347410f853f17f5d5f32d655a5066674~~","Cargo.toml":"~~7d7b91e75055586980888798514b6723844d2e6c69b77de85a56908341f129fa~~","Cargo.toml.orig":"~~0dbc03ef13c689b9ef343e031c2d4bd043bd59e862410644f4689a90ab539215~~","LICENSE-APACHE":"a60eea817514531668d7e00765731449fe14d059d3249e0bc93b36de45f759f2","LICENSE-MIT":"86dd7f026f916daf7511e39951ad8ea8cf55a8db67ae64060dacf829761c18f3","README.md":"0f0989d8abaf2539dc05d6b0dc5851dc432c6badae34d9bd23b76c56991ea254","benches/html5ever.rs":"848ca1aac420dd2dd9a7ca19e3fc9ec7737b47e5ea328e5021688eba5270066a","data/bench/lipsum-zh.html":"fafe3ed44b07f9cdb0a3ca5c010a3a73cc2534b64605593e71cc70a8cf8e27e4","data/bench/lipsum.html":"f37900ddb4ee463b9aa3191297bc1a4320b94bc72ee630de01882b50eb241740","data/bench/medium-fragment.html":"d21ab12c5e538ae48af6df7ee7d7939c6fe3b2d1e36b5874a3bbb957847ddba0","data/bench/small-fragment.html":"3825226a96ac919b0a69ae98e8547a5430214b16e56497570d382720af2bfca3","data/bench/strong.html":"5adb31981cca062df929353c60dfa1c7fb91170b5a389a13c8be6555a4cf107d","data/bench/tiny-fragment.html":"9cc2d58507945020d4d206738c8f166a618dc15161609935ab1975b913967edf","examples/arena.rs":"~~f3f8ad83b3e7a96fc826d0b95c161cdeb8aa12c86f9655310fabb04ca59ec3c5~~","examples/~~capi/tokenize.c":"ca0d98155c275f0d25ef7ce6735dec9d7cedb89f54416a406f53f10c616656bf","examples/~~noop-tokenize.rs":"79c1d34c8c941e3e6b6a1953455d0502fdc56195a7d932cdd7665a0011a9da9a","examples/noop-tree-builder.rs":"c87386e9f4b83cd52d7ffa8ff22cc0dd54f5f0b54398e7d41fb7a6352d9bfa72","examples/print-tree-actions.rs":"d04287c1a129fe7c18ff621cfe1a69c5411fa905733f895af420d948e80ec65d","examples/tokenize.rs":"fbf13646df42d4536691c62422b073e80b3f5cacfbc47294130c265134e21337","src/driver.rs":"~~efc2cfdfe6920fb9d4db71cc830bfbcebea53761f95562870601f3ec5ef83295~~","src/lib.rs":"~~8c0886497cd04c1e03d27454b7e61c74f5daf301865d1d21a2a1231b09206ca9~~","src/macros.rs":"9d964d34792195764fc7d53c7bfdec3ba468ee7708ab4dd0891a37c211f96977","src/serialize/mod.rs":"~~4590741ad63e40e624582ce68ebcb8e9b27cf11c3d943e68a974c5b481a205ca~~","src/tokenizer/char_ref/mod.rs":"e753058cb7ab64c63bb13e2156c583bdc60a8468efe6b9c64b8768c6effadf8f","src/tokenizer/interface.rs":"~~3c96dd965b492de373347c6a4d1ee2df58c319823ac6c2cfed3f20c9030f341e~~","src/tokenizer/mod.rs":"~~78f5739f536e65c711bf47e8b5760b005ddc9552d5316b5654c3e1fdeec4fae3~~","src/tokenizer/states.rs":"~~09a8d3c36072e8dc584fd0d520c2e250d8d7b1b10b3c8c90c388346f0795d78a~~","src/tree_builder/data.rs":"256c72e7d368b0be3d1f473c7b36907b6f6b602aff7d1441b9be109fea3375b1","src/tree_builder/mod.rs":"~~5c55e9a511d413f64329be2969b9f055126091e60e80837bdba7b89009b300ff~~","src/tree_builder/rules.rs":"~~8f275a7925d8592e351eee3e7f227bb88d48c65b91b9fc1bfb93dca722b1309d~~","src/tree_builder/tag_sets.rs":"~~3b70c6f8d86d06b0cb126d5a11541064766af324498752ca134909683d50ff30~~","src/tree_builder/types.rs":"~~dfef44b2ede374be5225d36a397524e3916dbf42232a6ded80299dd55bb91f80~~","src/util/str.rs":"48ca91feaa5e352cafb1743f996f57e6209c34851da6400488fb2fede2593161"},"package":"~~6452c4751a24e1b99c3260d505eaeee76a050573e61f30ac2c924ddc7236f01e~~"}
1	+ {"files":{".cargo_vcs_info.json":"4c1684373bc1b782c1a9e59ccb80b2a8640ac6e94840b9c23976c1cbf2e9c33f","Cargo.lock":"68a8d3b7a9b17cab79e4356bfc22638854725eb15231c51b415a8b63e7e1a569","Cargo.toml":"6482ec4d7aefb3bd9fbbe89ea619091114c3d566a309fc940f008c1a781ef213","Cargo.toml.orig":"512759192eb48ad87acc334ec5284db60cc9e5d461e551a1aa7c7fbe53b140f1","LICENSE-APACHE":"a60eea817514531668d7e00765731449fe14d059d3249e0bc93b36de45f759f2","LICENSE-MIT":"86dd7f026f916daf7511e39951ad8ea8cf55a8db67ae64060dacf829761c18f3","README.md":"0f0989d8abaf2539dc05d6b0dc5851dc432c6badae34d9bd23b76c56991ea254","benches/html5ever.rs":"848ca1aac420dd2dd9a7ca19e3fc9ec7737b47e5ea328e5021688eba5270066a","data/bench/lipsum-zh.html":"fafe3ed44b07f9cdb0a3ca5c010a3a73cc2534b64605593e71cc70a8cf8e27e4","data/bench/lipsum.html":"f37900ddb4ee463b9aa3191297bc1a4320b94bc72ee630de01882b50eb241740","data/bench/medium-fragment.html":"d21ab12c5e538ae48af6df7ee7d7939c6fe3b2d1e36b5874a3bbb957847ddba0","data/bench/small-fragment.html":"3825226a96ac919b0a69ae98e8547a5430214b16e56497570d382720af2bfca3","data/bench/strong.html":"5adb31981cca062df929353c60dfa1c7fb91170b5a389a13c8be6555a4cf107d","data/bench/tiny-fragment.html":"9cc2d58507945020d4d206738c8f166a618dc15161609935ab1975b913967edf","examples/arena.rs":"485dae462d5edfdf786f14018c525d214567877c61f5b4417a3ff6926b0a8e31","examples/noop-tokenize.rs":"79c1d34c8c941e3e6b6a1953455d0502fdc56195a7d932cdd7665a0011a9da9a","examples/noop-tree-builder.rs":"c87386e9f4b83cd52d7ffa8ff22cc0dd54f5f0b54398e7d41fb7a6352d9bfa72","examples/print-tree-actions.rs":"d04287c1a129fe7c18ff621cfe1a69c5411fa905733f895af420d948e80ec65d","examples/tokenize.rs":"fbf13646df42d4536691c62422b073e80b3f5cacfbc47294130c265134e21337","src/driver.rs":"760eca0d36fcb90c07858102928aa41efea9adb1eac2df74b972352c2b8c21b3","src/encoding.rs":"5b569fcd7f419fb521f86f970f53e943f5c23cc5afce32cd4364b66057c7bc42","src/lib.rs":"4b9193c2483083e52713aaa71689c18eeada669cbca3c1e08097a02141e2582b","src/macros.rs":"9d964d34792195764fc7d53c7bfdec3ba468ee7708ab4dd0891a37c211f96977","src/serialize/mod.rs":"b1c12517e34d4b109bdcecdb34895b6838f94b863c7f533e83f77c63aad2f1e2","src/tokenizer/char_ref/mod.rs":"e753058cb7ab64c63bb13e2156c583bdc60a8468efe6b9c64b8768c6effadf8f","src/tokenizer/interface.rs":"2aae4820e2d1061bc01df08cb1010fd2156beaf644390f0950bca56883444e8e","src/tokenizer/mod.rs":"008b5c3a6882266994710791d8926dd738d9e5472cac13ccfee2e90e575333a7","src/tokenizer/states.rs":"35ea888082a92b77a428af263224169b05ff5740bd519efdace90e44858a0b7d","src/tree_builder/data.rs":"256c72e7d368b0be3d1f473c7b36907b6f6b602aff7d1441b9be109fea3375b1","src/tree_builder/mod.rs":"70f653b2e9595b8c678873cc609b0c925e15f08e973fc31388ee763d915b24dd","src/tree_builder/rules.rs":"6eb715d41a62604dfbe8a6c46fa82efeab00cad1002fb40aab7c6e6f6dbdc0d2","src/tree_builder/tag_sets.rs":"5f1e1adffe46f4af0fe063d5847cf1e0fcc03312ca2e3b0062945a341b836716","src/tree_builder/types.rs":"85d2b3f8221faa9cbb561ad1dba8443ad4ed988b8a544511d3e01c28cdc8992f","src/util/str.rs":"48ca91feaa5e352cafb1743f996f57e6209c34851da6400488fb2fede2593161","tests/driver.rs":"ed917e950d7d43ade5e6eb63961bf40c77729d1df074d9f8cee5575daa9d2826"},"package":"1054432bae2f14e0061e33d23402fbaa67a921d319d56adc6bcf887ddad1cbc2"}

data/rust-vendor/html5ever/.cargo_vcs_info.json CHANGED Viewed

@@ -1,6 +1,6 @@
 {
   "git": {
-    "sha1": "34059007a855d284608f9447ed74c0c209bc0ad8"
+    "sha1": "be20a551d52b15f21fca6fa85f76e40d79f2eeb3"
   },
   "path_in_vcs": "html5ever"
 }