mdformat-space-control 0.4.0__tar.gz → 0.4.2__tar.gz

{mdformat_space_control-0.4.0 → mdformat_space_control-0.4.2}/CLAUDE.md

@@ -10,9 +10,10 @@ mdformat-space-control is an mdformat plugin that provides unified control over
 - **Frontmatter spacing**: Normalizes spacing after YAML frontmatter (works with mdformat-frontmatter)
 - **Consecutive blank line normalization**: Limits runs of 3+ empty lines to a maximum of 2
 - **Trailing whitespace removal**: Strips trailing whitespace outside code blocks
+- **Smart dash conversion**: Converts `--` to en-dash and `---` to em-dash, with protection for HTML comments, tags, code blocks, and inline code spans
 - **Escaped link repair**: Fixes malformed multi-line links from web-clipped content
 - **Wikilink preservation**: Handles Obsidian-style `[[links]]`, `[[links|aliases]]`, `[[page#heading]]`, `[[page#^blockid]]`, and `![[embeds]]`
-- **Explicit line breaks**: Converts soft breaks to hard breaks (`\` + newline) in paragraphs, list items, and blockquotes
+- **Soft break joining**: Joins soft breaks (plain newlines) into single lines in paragraphs, list items, and blockquotes
 
 This plugin merges functionality from mdformat-editorconfig and mdformat-tight-lists into a single plugin, solving the issue where mdformat only applies one set of list renderers when multiple plugins are installed.
 
@@ -41,11 +42,11 @@ mdformat_space_control/
 - **`plugin.py`**: Provides renderers, postprocessors, and parser extensions:
   - `_wikilink_rule`: Inline parser for Obsidian-style wikilinks
   - `_render_wikilink`: Preserves wikilinks unchanged
-  - `_render_softbreak`: Converts soft breaks to hard breaks (backslash + newline)
+  - `_render_softbreak`: Joins soft breaks (plain newlines) into single lines with spaces
   - `_render_list_item`: Per-item tight/loose formatting based on paragraph count
   - `_render_bullet_list`: Configurable indent + content-based tight/loose
   - `_render_ordered_list`: Configurable indent + content-based tight/loose
-  - `_postprocess_root`: Combined postprocessor applying frontmatter spacing, escaped link repair, consecutive blank line normalization, and trailing whitespace removal
+  - `_postprocess_root`: Combined postprocessor applying frontmatter spacing, smart dash conversion, escaped link repair, consecutive blank line normalization, and trailing whitespace removal
 
 ## Plugin Extension Points
 
@@ -66,6 +67,7 @@ space_control = "mdformat_space_control"
 - **`tests/test_fixtures.py`**: Parametrized fixture tests
 - **`tests/test_editorconfig.py`**: EditorConfig-specific tests using temp directories
 - **`tests/test_frontmatter.py`**: Frontmatter spacing tests (requires mdformat-frontmatter)
+- **`tests/test_dash_conversion.py`**: Smart dash conversion with HTML comment/tag protection
 - **`tests/test_spacing_features.py`**: Trailing whitespace, hard breaks, escaped link repair, consecutive blank line tests
 - **`tests/test_integration.py`**: Full-stack integration tests combining multiple features
 - **`tests/test_plugin_interactions.py`**: Tests for compatibility with other mdformat plugins

{mdformat_space_control-0.4.0 → mdformat_space_control-0.4.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mdformat_space_control
-Version: 0.4.0
+Version: 0.4.2
 Summary: An mdformat plugin for space control: EditorConfig indentation, tight lists, frontmatter spacing, smart dash conversion, explicit line breaks, and wikilinks.
 Project-URL: Homepage, https://github.com/jdmonaco/mdformat-space-control
 Author-email: Joseph Monaco <joe@selfmotion.net>
@@ -39,9 +39,9 @@ An [mdformat](https://github.com/executablebooks/mdformat) plugin that provides
 - **Consecutive blank line normalization**: Limits runs of 3+ empty lines to a maximum of 2
 - **Trailing whitespace removal**: Strips trailing whitespace outside code blocks
 - **Escaped link repair**: Fixes malformed multi-line links from web-clipped content
-- **Smart dash conversion**: Converts `--` to en-dash (–) and `---` to em-dash (—), preserving code blocks and inline code
+- **Smart dash conversion**: Converts `--` to en-dash (–) and `---` to em-dash (—), preserving code blocks, inline code, HTML comments, and HTML tags
 - **Wikilink preservation**: Handles Obsidian-style `[[links]]`, `[[links|aliases]]`, `[[page#heading]]`, `[[page#^blockid]]`, and `![[embeds]]`
-- **Explicit line breaks**: Converts soft breaks to hard breaks (`\` + newline) so source line breaks render visually in all Markdown renderers
+- **Soft break joining**: Joins soft breaks (plain newlines within paragraphs) into single lines, normalizing to single-line paragraphs
 
 ## Installation
 
@@ -193,7 +193,7 @@ The result—unexpected as it was—changed everything.
 Pages 10–20 of the report.
 ```
 
-Dashes are preserved inside fenced code blocks and inline code spans. Thematic breaks (`---`) and frontmatter delimiters are not affected. Sequences of 4+ dashes are left unchanged.
+Dashes are preserved inside fenced code blocks, inline code spans, HTML comments, and HTML tags. Thematic breaks (`---`) and frontmatter delimiters are not affected. Sequences of 4+ dashes are left unchanged.
 
 ### Wikilink Preservation
 
@@ -212,9 +212,9 @@ Wikilinks inside markdown link text are correctly handled without duplication:
 [![[image.jpg]]](http://example.com)
 ```
 
-### Explicit Line Breaks
+### Soft Break Joining
 
-Soft breaks (plain newlines within paragraphs) are converted to hard breaks (backslash + newline). This makes source line breaks explicit so they render visually regardless of the Markdown renderer's line-break strictness setting. The conversion applies to paragraphs, list items, and blockquotes.
+Soft breaks (plain newlines within paragraphs) are joined into single lines with spaces. This normalizes editor-inserted line wraps to single-line paragraphs, matching CommonMark rendering behavior where soft breaks produce spaces in HTML output. The joining applies to paragraphs, list items, and blockquotes.
 
 **Before:**
 ```markdown
@@ -224,14 +224,10 @@ a soft line break in the source.
 
 **After:**
 ```markdown
-This is a paragraph with\
-a hard line break in the output.
+This is a paragraph with a soft line break in the source.
 ```
 
-> **Note:** Without a GFM table parser, markdown-it treats pipe tables as paragraphs
-> containing soft breaks, which causes backslashes to appear at the end of table rows.
-> Install `mdformat-gfm` to enable proper table parsing and avoid this issue
-> (see [Compatible Plugins](#compatible-plugins)).
+Explicit hard breaks (`\` + newline) are preserved unchanged.
 
 ## Compatible Plugins
 

{mdformat_space_control-0.4.0 → mdformat_space_control-0.4.2}/README.md

@@ -11,9 +11,9 @@ An [mdformat](https://github.com/executablebooks/mdformat) plugin that provides
 - **Consecutive blank line normalization**: Limits runs of 3+ empty lines to a maximum of 2
 - **Trailing whitespace removal**: Strips trailing whitespace outside code blocks
 - **Escaped link repair**: Fixes malformed multi-line links from web-clipped content
-- **Smart dash conversion**: Converts `--` to en-dash (–) and `---` to em-dash (—), preserving code blocks and inline code
+- **Smart dash conversion**: Converts `--` to en-dash (–) and `---` to em-dash (—), preserving code blocks, inline code, HTML comments, and HTML tags
 - **Wikilink preservation**: Handles Obsidian-style `[[links]]`, `[[links|aliases]]`, `[[page#heading]]`, `[[page#^blockid]]`, and `![[embeds]]`
-- **Explicit line breaks**: Converts soft breaks to hard breaks (`\` + newline) so source line breaks render visually in all Markdown renderers
+- **Soft break joining**: Joins soft breaks (plain newlines within paragraphs) into single lines, normalizing to single-line paragraphs
 
 ## Installation
 
@@ -165,7 +165,7 @@ The result—unexpected as it was—changed everything.
 Pages 10–20 of the report.
 ```
 
-Dashes are preserved inside fenced code blocks and inline code spans. Thematic breaks (`---`) and frontmatter delimiters are not affected. Sequences of 4+ dashes are left unchanged.
+Dashes are preserved inside fenced code blocks, inline code spans, HTML comments, and HTML tags. Thematic breaks (`---`) and frontmatter delimiters are not affected. Sequences of 4+ dashes are left unchanged.
 
 ### Wikilink Preservation
 
@@ -184,9 +184,9 @@ Wikilinks inside markdown link text are correctly handled without duplication:
 [![[image.jpg]]](http://example.com)
 ```
 
-### Explicit Line Breaks
+### Soft Break Joining
 
-Soft breaks (plain newlines within paragraphs) are converted to hard breaks (backslash + newline). This makes source line breaks explicit so they render visually regardless of the Markdown renderer's line-break strictness setting. The conversion applies to paragraphs, list items, and blockquotes.
+Soft breaks (plain newlines within paragraphs) are joined into single lines with spaces. This normalizes editor-inserted line wraps to single-line paragraphs, matching CommonMark rendering behavior where soft breaks produce spaces in HTML output. The joining applies to paragraphs, list items, and blockquotes.
 
 **Before:**
 ```markdown
@@ -196,14 +196,10 @@ a soft line break in the source.
 
 **After:**
 ```markdown
-This is a paragraph with\
-a hard line break in the output.
+This is a paragraph with a soft line break in the source.
 ```
 
-> **Note:** Without a GFM table parser, markdown-it treats pipe tables as paragraphs
-> containing soft breaks, which causes backslashes to appear at the end of table rows.
-> Install `mdformat-gfm` to enable proper table parsing and avoid this issue
-> (see [Compatible Plugins](#compatible-plugins)).
+Explicit hard breaks (`\` + newline) are preserved unchanged.
 
 ## Compatible Plugins
 

{mdformat_space_control-0.4.0 → mdformat_space_control-0.4.2}/mdformat_space_control/__init__.py

@@ -1,6 +1,6 @@
-"""An mdformat plugin for space control: EditorConfig indentation, tight lists, frontmatter spacing, smart dash conversion, explicit line breaks, and wikilinks."""
+"""An mdformat plugin for space control: EditorConfig indentation, tight lists, frontmatter spacing, smart dash conversion, soft break joining, and wikilinks."""
 
-__version__ = "0.4.0"
+__version__ = "0.4.2"
 
 from .config import (
     get_current_file,

{mdformat_space_control-0.4.0 → mdformat_space_control-0.4.2}/mdformat_space_control/plugin.py

@@ -274,13 +274,14 @@ def _render_wikilink(node: RenderTreeNode, context: RenderContext) -> str:
 
 
 def _render_softbreak(node: RenderTreeNode, context: RenderContext) -> str:
-    """Render soft breaks with terminal backslash.
+    """Join soft breaks into a single line.
 
-    Converts soft breaks (plain newlines within paragraphs) to hard breaks
-    (backslash + newline) to preserve visible line-break rendering and avoid
-    relying on trailing whitespace.
+    Soft breaks (plain newlines within paragraphs) are editor-inserted
+    line wraps, not intentional visible breaks. Join them with a space
+    to normalize to single-line paragraphs. This matches CommonMark
+    rendering behavior where soft breaks produce spaces in HTML output.
     """
-    return "\\" + "\n"
+    return " "
 
 
 # A mapping from syntax tree node type to a function that renders it.
@@ -391,7 +392,8 @@ def _convert_dash_sequences(text: str) -> str:
     """Convert markdown dash sequences to Unicode em-dash and en-dash.
 
     Converts ``---`` to em-dash (U+2014) and ``--`` to en-dash (U+2013).
-    Preserves dashes inside fenced code blocks and inline code spans.
+    Preserves dashes inside fenced code blocks, inline code spans,
+    HTML comments, and HTML tags.
     Skips lines that are only dashes (thematic breaks, frontmatter delimiters).
 
     Em-dash is matched first (longer sequence) to prevent ``---`` from being
@@ -400,6 +402,10 @@ def _convert_dash_sequences(text: str) -> str:
     # Patterns for inline code span placeholders
     inline_code_re = re.compile(r"(`+)(.+?)\1")
 
+    # HTML protection patterns
+    html_comment_re = re.compile(r"<!--.*?-->")
+    html_tag_re = re.compile(r"<[^>]+>")
+
     # Dash conversion patterns: negative lookahead/behind prevent matching
     # 4+ dash sequences (e.g., ---- horizontal rules in some contexts)
     em_dash_re = re.compile(r"(?<!-)---(?!-)")
@@ -411,6 +417,7 @@ def _convert_dash_sequences(text: str) -> str:
     lines = text.split("\n")
     result = []
     in_code_block = False
+    in_html_comment = False
 
     for line in lines:
         stripped = line.lstrip()
@@ -419,23 +426,38 @@ def _convert_dash_sequences(text: str) -> str:
             result.append(line)
         elif in_code_block:
             result.append(line)
+        elif in_html_comment:
+            # Inside a multi-line HTML comment, skip dash conversion
+            if "-->" in line:
+                in_html_comment = False
+            result.append(line)
         elif only_dashes_re.match(line):
             result.append(line)
         else:
-            # Protect inline code spans with placeholders
+            # Check for multi-line HTML comment opening (no closing on same line)
+            if "<!--" in line and "-->" not in line:
+                in_html_comment = True
+                result.append(line)
+                continue
+
+            # Protect inline code spans and HTML elements with placeholders
             placeholders = []
 
-            def _save_code_span(m: re.Match) -> str:
+            def _save_placeholder(m: re.Match) -> str:
                 placeholders.append(m.group(0))
                 return f"\x00CODE{len(placeholders) - 1}\x00"
 
-            protected = inline_code_re.sub(_save_code_span, line)
+            protected = inline_code_re.sub(_save_placeholder, line)
+
+            # Protect HTML comments first (longer matches), then tags
+            protected = html_comment_re.sub(_save_placeholder, protected)
+            protected = html_tag_re.sub(_save_placeholder, protected)
 
             # Convert em-dash first (longer match), then en-dash
             protected = em_dash_re.sub("\u2014", protected)
             protected = en_dash_re.sub("\u2013", protected)
 
-            # Restore inline code spans
+            # Restore placeholders
             for i, original in enumerate(placeholders):
                 protected = protected.replace(f"\x00CODE{i}\x00", original)