html-to-markdown 2.6.0__tar.gz → 2.6.2__tar.gz

{html_to_markdown-2.6.0 → html_to_markdown-2.6.2}/Cargo.lock

@@ -997,7 +997,7 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-cli"
-version = "2.6.0"
+version = "2.6.2"
 dependencies = [
  "assert_cmd",
  "clap",
@@ -1011,7 +1011,7 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-node"
-version = "2.6.0"
+version = "2.6.2"
 dependencies = [
  "html-to-markdown-rs",
  "mimalloc-rust",
@@ -1022,7 +1022,7 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-php"
-version = "2.6.0"
+version = "2.6.2"
 dependencies = [
  "ext-php-rs",
  "html-to-markdown-rs",
@@ -1030,7 +1030,7 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-py"
-version = "2.6.0"
+version = "2.6.2"
 dependencies = [
  "base64",
  "html-to-markdown-rs",
@@ -1049,7 +1049,7 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-rs"
-version = "2.6.0"
+version = "2.6.2"
 dependencies = [
  "ammonia",
  "base64",
@@ -1066,7 +1066,7 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-wasm"
-version = "2.6.0"
+version = "2.6.2"
 dependencies = [
  "console_error_panic_hook",
  "html-to-markdown-rs",

{html_to_markdown-2.6.0 → html_to_markdown-2.6.2}/Cargo.toml

@@ -3,7 +3,7 @@ resolver = "2"
 members = ["crates/html-to-markdown-py"]
 
 [workspace.package]
-version = "2.6.0"
+version = "2.6.2"
 edition = "2021"
 authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]
 license = "MIT"
@@ -15,7 +15,7 @@ rust-version = "1.80"
 
 [workspace.dependencies]
 # Core library
-html-to-markdown-rs = { version = "2.6.0", path = "crates/html-to-markdown" }
+html-to-markdown-rs = { version = "2.6.1", path = "crates/html-to-markdown" }
 
 # HTML parsing and sanitization
 tl = "0.7"

{html_to_markdown-2.6.0 → html_to_markdown-2.6.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: html-to-markdown
-Version: 2.6.0
+Version: 2.6.2
 Classifier: Development Status :: 5 - Production/Stable
 Classifier: Environment :: Console
 Classifier: Intended Audience :: Developers

{html_to_markdown-2.6.0 → html_to_markdown-2.6.2}/crates/html-to-markdown/benches/conversion_benchmark.rs

@@ -3,8 +3,9 @@
 //! Measures throughput (ops/sec, MB/sec) and provides performance baselines
 //! for the core Rust conversion engine.
 
-use criterion::{black_box, criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
+use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion, Throughput};
 use html_to_markdown_rs::{convert, CodeBlockStyle, ConversionOptions, HeadingStyle};
+use std::hint::black_box;
 use std::time::Duration;
 
 /// Generate HTML with varying complexity

{html_to_markdown-2.6.0 → html_to_markdown-2.6.2}/crates/html-to-markdown/benches/micro_benchmark.rs

@@ -1,7 +1,8 @@
 //! Micro-benchmarks for specific operations
 
-use criterion::{black_box, criterion_group, criterion_main, BenchmarkId, Criterion};
+use criterion::{criterion_group, criterion_main, BenchmarkId, Criterion};
 use html_to_markdown_rs::{convert, ConversionOptions};
+use std::hint::black_box;
 
 /// Benchmark text-heavy documents
 fn bench_text_operations(c: &mut Criterion) {

{html_to_markdown-2.6.0 → html_to_markdown-2.6.2}/crates/html-to-markdown/src/converter.rs

@@ -4522,7 +4522,7 @@ fn convert_table_row(
     options: &ConversionOptions,
     ctx: &Context,
     row_index: usize,
-    rowspan_tracker: &mut std::collections::HashMap<usize, usize>,
+    rowspan_tracker: &mut std::collections::HashMap<usize, (String, usize)>,
     dom_ctx: &DomContext,
 ) {
     let mut row_text = String::with_capacity(256);
@@ -4546,8 +4546,10 @@ fn convert_table_row(
     let mut cell_iter = cells.iter();
 
     loop {
-        if let Some(remaining_rows) = rowspan_tracker.get_mut(&col_index) {
+        if let Some((content, remaining_rows)) = rowspan_tracker.get_mut(&col_index) {
             if *remaining_rows > 0 {
+                row_text.push(' ');
+                row_text.push_str(content);
                 row_text.push_str(" |");
                 *remaining_rows -= 1;
                 if *remaining_rows == 0 {
@@ -4559,12 +4561,21 @@ fn convert_table_row(
         }
 
         if let Some(cell_handle) = cell_iter.next() {
+            let cell_start = row_text.len();
             convert_table_cell(cell_handle, parser, &mut row_text, options, ctx, "", dom_ctx);
 
             let (colspan, rowspan) = get_colspan_rowspan(cell_handle, parser);
 
             if rowspan > 1 {
-                rowspan_tracker.insert(col_index, rowspan - 1);
+                // Extract the cell content that was just added (without separators)
+                let cell_text = &row_text[cell_start..];
+                // Strip leading space and trailing " |"
+                let cell_content = cell_text
+                    .trim_start_matches(' ')
+                    .trim_end_matches(" |")
+                    .trim()
+                    .to_string();
+                rowspan_tracker.insert(col_index, (cell_content, rowspan - 1));
             }
 
             col_index += colspan;
@@ -4920,8 +4931,10 @@ mod tests {
     #[test]
     fn test_preserve_tags_simple_table() {
         let html = r#"<div><table><tr><td>Cell 1</td><td>Cell 2</td></tr></table><p>Text</p></div>"#;
-        let mut options = ConversionOptions::default();
-        options.preserve_tags = vec!["table".to_string()];
+        let options = ConversionOptions {
+            preserve_tags: vec!["table".to_string()],
+            ..Default::default()
+        };
         let result = convert_html(html, &options).unwrap();
 
         assert!(result.contains("<table>"), "Should preserve table tag");
@@ -4934,8 +4947,10 @@ mod tests {
     #[test]
     fn test_preserve_tags_with_attributes() {
         let html = r#"<table class="data" id="mytable"><tr><td>Data</td></tr></table>"#;
-        let mut options = ConversionOptions::default();
-        options.preserve_tags = vec!["table".to_string()];
+        let options = ConversionOptions {
+            preserve_tags: vec!["table".to_string()],
+            ..Default::default()
+        };
         let result = convert_html(html, &options).unwrap();
 
         assert!(result.contains("<table"), "Should preserve table tag");
@@ -4947,8 +4962,10 @@ mod tests {
     #[test]
     fn test_preserve_tags_multiple_tags() {
         let html = r#"<div><table><tr><td>Table</td></tr></table><form><input type="text"/></form><p>Text</p></div>"#;
-        let mut options = ConversionOptions::default();
-        options.preserve_tags = vec!["table".to_string(), "form".to_string()];
+        let options = ConversionOptions {
+            preserve_tags: vec!["table".to_string(), "form".to_string()],
+            ..Default::default()
+        };
         let result = convert_html(html, &options).unwrap();
 
         assert!(result.contains("<table>"), "Should preserve table");
@@ -4959,8 +4976,10 @@ mod tests {
     #[test]
     fn test_preserve_tags_nested_content() {
         let html = r#"<table><thead><tr><th>Header</th></tr></thead><tbody><tr><td>Data</td></tr></tbody></table>"#;
-        let mut options = ConversionOptions::default();
-        options.preserve_tags = vec!["table".to_string()];
+        let options = ConversionOptions {
+            preserve_tags: vec!["table".to_string()],
+            ..Default::default()
+        };
         let result = convert_html(html, &options).unwrap();
 
         assert!(result.contains("<thead>"), "Should preserve nested thead");
@@ -4985,9 +5004,11 @@ mod tests {
     #[test]
     fn test_preserve_tags_vs_strip_tags() {
         let html = r#"<table><tr><td>Table</td></tr></table><div><span>Text</span></div>"#;
-        let mut options = ConversionOptions::default();
-        options.preserve_tags = vec!["table".to_string()];
-        options.strip_tags = vec!["span".to_string()];
+        let options = ConversionOptions {
+            preserve_tags: vec!["table".to_string()],
+            strip_tags: vec!["span".to_string()],
+            ..Default::default()
+        };
         let result = convert_html(html, &options).unwrap();
 
         assert!(result.contains("<table>"), "Should preserve table");

{html_to_markdown-2.6.0 → html_to_markdown-2.6.2}/pyproject.toml

@@ -7,7 +7,7 @@ requires = [
 
 [project]
 name = "html-to-markdown"
-version = "2.6.0"
+version = "2.6.2"
 description = "High-performance HTML to Markdown converter powered by Rust with a clean Python API"
 readme = "README.md"
 keywords = [