html-to-markdown 3.0.1 → 3.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7681da88f184d981337443e5fd5ba46746c3d51690aea94cd4ed33a903081996
-  data.tar.gz: e0073d5087e894e76588d03853f94d618412e650d8d3c6b62831555ede8f5230
+  metadata.gz: e2761dc167e2c7f7e0e27da4367660d7dc4d18f853b5b5add976019434a37de0
+  data.tar.gz: 8d6822eb08fc782524c4ec35446c48ff017bd2aa640f2515e79c05e4d89508b7
 SHA512:
-  metadata.gz: 94d123fc5f89d7d83ed6e05feed6fe9d6a084822d42c257a41e2af2bfdde92905f852956dea05e924483284b1fde84eaf1b0e0230ab542c5db56fd64d5a553bf
-  data.tar.gz: e474ddd110717cada39806796d8c7b1a2bd8900193d4413058392f69424ca7a964c3fa9b41feed493526ccbb6d7f3a6c36d3cc0553504c28c6933488abdbe9a9
+  metadata.gz: 93b26fafdae4c4beca9fc6134a3ba2948369210900c8145e1ad72b1714db86e453523361e39308b8d78d5b94d8dc83e33259c2ae2eca727c1b20f3c1629b81c7
+  data.tar.gz: 69d16a9ff3a3d67a3cd267d9c0ff7ed48b9f3736e01fa87f5ec7ae2dd1b79263dea40d9e54392ab50546d12ccfd0c9486b3a728e45344009769a1f1fd2a65bd2

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    html-to-markdown (3.0.1)
+    html-to-markdown (3.0.2)
       rb_sys (>= 0.9, < 1.0)
 
 GEM
@@ -29,6 +29,7 @@ GEM
     diff-lcs (1.6.2)
     drb (2.2.3)
     ffi (1.17.4-arm64-darwin)
+    ffi (1.17.4-x86_64-linux-gnu)
     fileutils (1.8.0)
     i18n (1.14.8)
       concurrent-ruby (~> 1.0)
@@ -40,7 +41,7 @@ GEM
       rb-fsevent (~> 0.10, >= 0.10.3)
       rb-inotify (~> 0.9, >= 0.9.10)
     logger (1.7.0)
-    minitest (6.0.2)
+    minitest (6.0.3)
       drb (~> 2.0)
       prism (~> 1.5)
     mutex_m (0.3.0)
@@ -58,7 +59,8 @@ GEM
     rb-fsevent (0.11.2)
     rb-inotify (0.11.1)
       ffi (~> 1.0)
-    rb_sys (0.9.124)
+    rb_sys (0.9.125)
+      json (>= 2)
       rake-compiler-dock (= 1.11.0)
     rbs (3.10.4)
       logger
@@ -126,6 +128,7 @@ GEM
 
 PLATFORMS
   arm64-darwin
+  x86_64-linux
 
 DEPENDENCIES
   html-to-markdown!
@@ -148,15 +151,16 @@ CHECKSUMS
   diff-lcs (1.6.2) sha256=9ae0d2cba7d4df3075fe8cd8602a8604993efc0dfa934cff568969efb1909962
   drb (2.2.3) sha256=0b00d6fdb50995fe4a45dea13663493c841112e4068656854646f418fda13373
   ffi (1.17.4-arm64-darwin) sha256=19071aaf1419251b0a46852abf960e77330a3b334d13a4ab51d58b31a937001b
+  ffi (1.17.4-x86_64-linux-gnu) sha256=9d3db14c2eae074b382fa9c083fe95aec6e0a1451da249eab096c34002bc752d
   fileutils (1.8.0) sha256=8c6b1df54e2540bdb2f39258f08af78853aa70bad52b4d394bbc6424593c6e02
-  html-to-markdown (3.0.1)
+  html-to-markdown (3.0.2)
   i18n (1.14.8) sha256=285778639134865c5e0f6269e0b818256017e8cde89993fdfcbfb64d088824a5
   json (2.19.3) sha256=289b0bb53052a1fa8c34ab33cc750b659ba14a5c45f3fcf4b18762dc67c78646
   language_server-protocol (3.17.0.5) sha256=fd1e39a51a28bf3eec959379985a72e296e9f9acfce46f6a79d31ca8760803cc
   lint_roller (1.1.0) sha256=2c0c845b632a7d172cb849cc90c1bce937a28c5c8ccccb50dfd46a485003cc87
   listen (3.10.0) sha256=c6e182db62143aeccc2e1960033bebe7445309c7272061979bb098d03760c9d2
   logger (1.7.0) sha256=196edec7cc44b66cfb40f9755ce11b392f21f7967696af15d274dde7edff0203
-  minitest (6.0.2) sha256=db6e57956f6ecc6134683b4c87467d6dd792323c7f0eea7b93f66bd284adbc3d
+  minitest (6.0.3) sha256=88ac8a1de36c00692420e7cb3cc11a0773bbcb126aee1c249f320160a7d11411
   mutex_m (0.3.0) sha256=cfcb04ac16b69c4813777022fdceda24e9f798e48092a2b817eb4c0a782b0751
   parallel (1.27.0) sha256=4ac151e1806b755fb4e2dc2332cbf0e54f2e24ba821ff2d3dcf86bf6dc4ae130
   parser (3.3.11.1) sha256=d17ace7aabe3e72c3cc94043714be27cc6f852f104d81aa284c2281aecc65d54
@@ -168,7 +172,7 @@ CHECKSUMS
   rake-compiler-dock (1.11.0) sha256=eab51f2cd533eb35cea6b624a75281f047123e70a64c58b607471bb49428f8c2
   rb-fsevent (0.11.2) sha256=43900b972e7301d6570f64b850a5aa67833ee7d87b458ee92805d56b7318aefe
   rb-inotify (0.11.1) sha256=a0a700441239b0ff18eb65e3866236cd78613d6b9f78fea1f9ac47a85e47be6e
-  rb_sys (0.9.124) sha256=513476557b12eaf73764b3da9f8746024558fe8699bda785fb548c9aa3877ae7
+  rb_sys (0.9.125) sha256=14efd4e07eaf7c07edb1bab548d2a4767869a47a8821fc5ea52d9bf982ef00a8
   rbs (3.10.4) sha256=b17d7c4be4bb31a11a3b529830f0aa206a807ca42f2e7921a3027dfc6b7e5ce8
   regexp_parser (2.11.3) sha256=ca13f381a173b7a93450e53459075c9b76a10433caadcb2f1180f2c741fc55a4
   rspec (3.13.2) sha256=206284a08ad798e61f86d7ca3e376718d52c0bc944626b2349266f239f820587

data/README.md

@@ -18,7 +18,7 @@
     <img src="https://img.shields.io/maven-central/v/dev.kreuzberg/html-to-markdown?label=Java&color=007ec6" alt="Java">
   </a>
   <a href="https://pkg.go.dev/github.com/kreuzberg-dev/html-to-markdown/packages/go/v3/htmltomarkdown">
-    <img src="https://img.shields.io/github/v/tag/kreuzberg-dev/html-to-markdown?label=Go&color=007ec6&filter=v3.0.1" alt="Go">
+    <img src="https://img.shields.io/github/v/tag/kreuzberg-dev/html-to-markdown?label=Go&color=007ec6&filter=v3.0.2" alt="Go">
   </a>
   <a href="https://www.nuget.org/packages/KreuzbergDev.HtmlToMarkdown/">
     <img src="https://img.shields.io/nuget/v/KreuzbergDev.HtmlToMarkdown?label=C%23&color=007ec6" alt="C#">

data/ext/html-to-markdown-rb/native/Cargo.lock

@@ -59,16 +59,14 @@ checksum = "72b3254f16251a8381aa12e40e3c4d2f0199f8c6508fbecb9d91f575e0fbb8c6"
 
 [[package]]
 name = "bindgen"
-version = "0.69.5"
+version = "0.72.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "271383c67ccabffb7381723dea0672a673f292304fcb45c01cc648c7a8d58088"
+checksum = "993776b509cfb49c750f11b8f07a46fa23e0a1386ffc01fb1e7d343efc387895"
 dependencies = [
  "bitflags",
  "cexpr",
  "clang-sys",
  "itertools",
- "lazy_static",
- "lazycell",
  "proc-macro2",
  "quote",
  "regex",
@@ -260,7 +258,7 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-rb"
-version = "3.0.0"
+version = "3.0.1"
 dependencies = [
  "html-to-markdown-rs",
  "magnus",
@@ -269,7 +267,7 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-rs"
-version = "3.0.0"
+version = "3.0.1"
 dependencies = [
  "ahash",
  "astral-tl",
@@ -326,9 +324,9 @@ dependencies = [
 
 [[package]]
 name = "itertools"
-version = "0.12.1"
+version = "0.13.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ba291022dbbd398a455acf126c1e341954079855bc60dfdda641363bd6922569"
+checksum = "413ee7dfc52ee1a4949ceeb7dbc8a33f2d6c088194d9f922fb8318faf1f01186"
 dependencies = [
  "either",
 ]
@@ -345,12 +343,6 @@ version = "1.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe"
 
-[[package]]
-name = "lazycell"
-version = "1.3.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
-
 [[package]]
 name = "libc"
 version = "0.2.183"
@@ -609,18 +601,18 @@ dependencies = [
 
 [[package]]
 name = "rb-sys"
-version = "0.9.124"
+version = "0.9.125"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c85c4188462601e2aa1469def389c17228566f82ea72f137ed096f21591bc489"
+checksum = "85b37650fabd8ba515910a0dc089dcb6348eb3c35fbf91698cb226435be2babc"
 dependencies = [
  "rb-sys-build",
 ]
 
 [[package]]
 name = "rb-sys-build"
-version = "0.9.124"
+version = "0.9.125"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "568068db4102230882e6d4ae8de6632e224ca75fe5970f6e026a04e91ed635d3"
+checksum = "c73b806faa66006e491458b48a78725621c1ac5a2a6efe2614c90711a7780b80"
 dependencies = [
  "bindgen",
  "lazy_static",
@@ -677,9 +669,9 @@ checksum = "dc897dd8d9e8bd1ed8cdad82b5966c3e0ecae09fb1907d58efaa013543185d0a"
 
 [[package]]
 name = "rustc-hash"
-version = "1.1.0"
+version = "2.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2"
+checksum = "94300abf3f1ae2e2b8ffb7b58043de3d399c73fa6f4b73826402a5c457614dbe"
 
 [[package]]
 name = "scopeguard"

data/ext/html-to-markdown-rb/native/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "html-to-markdown-rb"
-version = "3.0.1"
+version = "3.0.2"
 edition = "2024"
 authors = ["Na'aman Hirschfeld <naaman@kreuzberg.dev>"]
 license = "MIT"

data/lib/html_to_markdown/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module HtmlToMarkdown
-  VERSION = '3.0.1'
+  VERSION = '3.0.2'
 end

data/vendor/Cargo.toml

@@ -3,7 +3,7 @@ members = ["html-to-markdown-rs"]
 resolver = "2"
 
 [workspace.package]
-version = "3.0.1"
+version = "3.0.2"
 edition = "2024"
 rust-version = "1.85"
 authors = ["Na'aman Hirschfeld <naaman@kreuzberg.dev>"]

data/vendor/html-to-markdown-rs/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "html-to-markdown-rs"
-version = "3.0.1"
+version = "3.0.2"
 edition = "2024"
 authors = ["Na'aman Hirschfeld <naaman@kreuzberg.dev>"]
 license = "MIT"

data/vendor/html-to-markdown-rs/src/converter/block/heading.rs

@@ -128,16 +128,20 @@ pub(crate) fn handle(
         }
 
         // Notify the structure collector if present.
-        if let Some(ref sc) = ctx.structure_collector {
-            if let Some(node) = node_handle.get(parser) {
-                if let tl::Node::Tag(tag) = node {
-                    let id = tag
-                        .attributes()
-                        .get("id")
-                        .flatten()
-                        .map(|v| v.as_utf8_str().to_string());
-                    sc.borrow_mut()
-                        .push_heading(level as u8, normalized.as_ref(), id.as_deref());
+        // Skip headings inside table cells — they are part of the table content,
+        // not standalone structural headings.
+        if !ctx.in_table_cell {
+            if let Some(ref sc) = ctx.structure_collector {
+                if let Some(node) = node_handle.get(parser) {
+                    if let tl::Node::Tag(tag) = node {
+                        let id = tag
+                            .attributes()
+                            .get("id")
+                            .flatten()
+                            .map(|v| v.as_utf8_str().to_string());
+                        sc.borrow_mut()
+                            .push_heading(level as u8, normalized.as_ref(), id.as_deref());
+                    }
                 }
             }
         }

data/vendor/html-to-markdown-rs/src/converter/handlers/code_block.rs

@@ -301,6 +301,10 @@ pub fn handle_pre(
         {
             format_code_block(&processed_content, language.as_deref(), output, options, ctx);
         }
+
+        if let Some(ref sc) = ctx.structure_collector {
+            sc.borrow_mut().push_code(&processed_content, language.as_deref());
+        }
     }
 }
 

data/vendor/html-to-markdown-rs/src/converter/handlers/image.rs

@@ -192,6 +192,12 @@ pub fn handle_img(
             }
         }
     }
+
+    if let Some(ref sc) = ctx.structure_collector {
+        let src_opt = if src.is_empty() { None } else { Some(src.as_ref()) };
+        let alt_opt = if alt.is_empty() { None } else { Some(alt.as_ref()) };
+        sc.borrow_mut().push_image(src_opt, alt_opt);
+    }
 }
 
 /// Format an image as Markdown syntax.

data/vendor/html-to-markdown-rs/src/converter/list/item.rs

@@ -204,6 +204,8 @@ pub(crate) fn handle_li(
             }
         }
 
+        let item_start_pos = output.len();
+
         let children = tag.children();
         {
             for child_handle in children.top().iter() {
@@ -213,6 +215,18 @@ pub(crate) fn handle_li(
 
         trim_trailing_whitespace(output);
 
+        if !ctx.in_table_cell {
+            if let Some(ref sc) = ctx.structure_collector {
+                if item_start_pos <= output.len() && output.is_char_boundary(item_start_pos) {
+                    let rendered = &output[item_start_pos..];
+                    let content = rendered.trim();
+                    if !content.is_empty() {
+                        sc.borrow_mut().push_list_item(content);
+                    }
+                }
+            }
+        }
+
         #[cfg(feature = "visitor")]
         if let Some(ref visitor_handle) = ctx.visitor {
             use crate::visitor::{NodeContext, NodeType, VisitResult};

data/vendor/html-to-markdown-rs/src/converter/list/ordered.rs

@@ -107,6 +107,12 @@ pub(crate) fn handle_ol(
         }
     }
 
+    if !ctx.in_table_cell {
+        if let Some(ref sc) = ctx.structure_collector {
+            sc.borrow_mut().push_list_start(true);
+        }
+    }
+
     process_list_children(
         *node_handle,
         parser,
@@ -121,6 +127,12 @@ pub(crate) fn handle_ol(
         dom_ctx,
     );
 
+    if !ctx.in_table_cell {
+        if let Some(ref sc) = ctx.structure_collector {
+            sc.borrow_mut().push_list_end();
+        }
+    }
+
     add_nested_list_trailing_separator(output, ctx);
 
     #[cfg(feature = "visitor")]

data/vendor/html-to-markdown-rs/src/converter/list/unordered.rs

@@ -101,6 +101,12 @@ pub(crate) fn handle_ul(
         }
     }
 
+    if !ctx.in_table_cell {
+        if let Some(ref sc) = ctx.structure_collector {
+            sc.borrow_mut().push_list_start(false);
+        }
+    }
+
     process_list_children(
         *node_handle,
         parser,
@@ -115,6 +121,12 @@ pub(crate) fn handle_ul(
         dom_ctx,
     );
 
+    if !ctx.in_table_cell {
+        if let Some(ref sc) = ctx.structure_collector {
+            sc.borrow_mut().push_list_end();
+        }
+    }
+
     add_nested_list_trailing_separator(output, ctx);
 
     #[cfg(feature = "visitor")]

data/vendor/html-to-markdown-rs/src/types/structure_builder.rs

@@ -226,7 +226,8 @@ fn make_node_id(node_type: &str, text: &str, index: usize) -> String {
     let mut hasher = DefaultHasher::new();
     node_type.hash(&mut hasher);
     // Only hash a prefix of the text to keep cost bounded.
-    text[..text.len().min(64)].hash(&mut hasher);
+    let end = crate::converter::utility::content::floor_char_boundary(text, text.len().min(64));
+    text[..end].hash(&mut hasher);
     index.hash(&mut hasher);
     let digest = hasher.finish();
     format!("{node_type}-{digest:016x}")

data/vendor/html-to-markdown-rs/src/types/structure_collector.rs

@@ -347,7 +347,8 @@ impl StructureCollector {
 
         let mut hasher = DefaultHasher::new();
         node_type.hash(&mut hasher);
-        text[..text.len().min(64)].hash(&mut hasher);
+        let end = crate::converter::utility::content::floor_char_boundary(text, text.len().min(64));
+        text[..end].hash(&mut hasher);
         index.hash(&mut hasher);
         let digest = hasher.finish();
         format!("{node_type}-{digest:016x}")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: html-to-markdown
 version: !ruby/object:Gem::Version
-  version: 3.0.1
+  version: 3.0.2
 platform: ruby
 authors:
 - Na'aman Hirschfeld
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-03-31 00:00:00.000000000 Z
+date: 2026-04-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys