@joaodotwork/md-2-pdf 1.3.0 → 1.4.0

package/README.md

@@ -20,6 +20,9 @@ Most markdown-to-pdf converters struggle with diagrams or produce poorly formatt
   - Interactive, blue clickable links.
   - Automatic widow/orphan protection (keeps headers with content).
   - Wide tables wrap to fit the page — pipe tables with short separator dashes (`|---|---|`) get equal column widths and wrapping cells instead of overflowing.
+  - Long inline code (filenames, identifiers) wraps cleanly inside narrow table cells, breaking at `-`, `_`, `.`, and `/` instead of colliding with the next column.
+  - Code blocks wrap soft-overflow lines instead of running off the page (via `fvextra` `breaklines`), with a `↪` continuation marker.
+  - Unicode-safe monospace font (DejaVu Sans Mono) — box-drawing characters (`│ └── ├──`) and other non-ASCII glyphs render correctly in code blocks.
   - Customizable margins and geometry.
 - **Multi-Engine Support:** Automatically detects and uses the best available PDF engine (`xelatex`, `pdflatex`, `weasyprint`, or `wkhtmltopdf`).
 - **Batch Processing:** Convert single files or entire directories with one command.

package/filters/code-break.lua

@@ -0,0 +1,40 @@
+-- Make long inline code spans (`like-this-filename.json`, `vngrd_woodcut_v3`)
+-- wrappable inside narrow table cells.
+--
+-- Pandoc emits inline code as \texttt{...}, which has no break opportunities
+-- inside the word (no spaces, hyphenchar disabled). Long identifiers then
+-- overflow their column. We insert a zero-width \allowbreak after each
+-- hyphen / underscore / dot / slash, so TeX can break the line there if it
+-- needs to. The visible character stays put.
+
+local break_after = { ['-'] = true, ['_'] = true, ['.'] = true, ['/'] = true }
+
+local function latex_escape(c)
+  if c == '\\' then return '\\textbackslash{}'
+  elseif c == '{'  then return '\\{'
+  elseif c == '}'  then return '\\}'
+  elseif c == '$'  then return '\\$'
+  elseif c == '&'  then return '\\&'
+  elseif c == '%'  then return '\\%'
+  elseif c == '#'  then return '\\#'
+  elseif c == '_'  then return '\\_'
+  elseif c == '^'  then return '\\textasciicircum{}'
+  elseif c == '~'  then return '\\textasciitilde{}'
+  else return c
+  end
+end
+
+function Code(elem)
+  if FORMAT ~= 'latex' and FORMAT ~= 'pdf' then return nil end
+  local text = elem.text
+  if not text:find('[-_./]') then return nil end
+  local out = {}
+  for i = 1, #text do
+    local c = text:sub(i, i)
+    out[#out + 1] = latex_escape(c)
+    if break_after[c] then
+      out[#out + 1] = '\\allowbreak{}'
+    end
+  end
+  return pandoc.RawInline('latex', '\\texttt{' .. table.concat(out) .. '}')
+end

package/md_to_pdf.py

@@ -170,6 +170,7 @@ def convert_markdown_to_pdf(
     # Locate bundled Lua filters
     script_dir = Path(__file__).resolve().parent
     table_fit_filter = script_dir / 'filters' / 'table-fit.lua'
+    code_break_filter = script_dir / 'filters' / 'code-break.lua'
 
     # Convert to PDF using pandoc
     cmd = [
@@ -179,6 +180,7 @@ def convert_markdown_to_pdf(
         '--from=gfm',
         f'--pdf-engine={pdf_engine}',
         f'--lua-filter={table_fit_filter}',
+        f'--lua-filter={code_break_filter}',
         '-V', 'geometry:top=0.75in',
         '-V', 'geometry:bottom=1in',
         '-V', 'geometry:left=0.75in',
@@ -189,8 +191,19 @@ def convert_markdown_to_pdf(
         '-V', 'colorlinks=true',
         '-V', 'linkcolor=blue',
         '-V', 'urlcolor=blue',
-        '-V', 'header-includes=\\renewcommand{\\rule}[2]{\\vspace{0.5em}} \\widowpenalty=10000 \\clubpenalty=10000 \\brokenpenalty=10000 \\setlength{\\emergencystretch}{3em} \\usepackage{newunicodechar} \\newunicodechar{·}{\\textperiodcentered\\allowbreak} \\newunicodechar{•}{\\textbullet\\allowbreak}'
+        '-V', 'header-includes=\\renewcommand{\\rule}[2]{\\vspace{0.5em}} \\widowpenalty=10000 \\clubpenalty=10000 \\brokenpenalty=10000 \\setlength{\\emergencystretch}{3em} \\usepackage{newunicodechar} \\newunicodechar{·}{\\textperiodcentered\\allowbreak} \\newunicodechar{•}{\\textbullet\\allowbreak} \\usepackage{fvextra} \\DefineVerbatimEnvironment{Highlighting}{Verbatim}{breaklines,commandchars=\\\\\\{\\}} \\DefineVerbatimEnvironment{verbatim}{Verbatim}{breaklines}'
     ]
+
+    # Use a Unicode-rich monospace font when the engine supports it (fontspec).
+    # DejaVu Sans Mono ships with TeX Live and covers box-drawing characters
+    # (│ └── ├──), so ASCII-art trees in code blocks render correctly. Pass the
+    # font as a .ttf filename so fontspec resolves it via kpsewhich, which
+    # works regardless of fontconfig setup or TeX Live install path.
+    if pdf_engine in ('xelatex', 'lualatex'):
+        cmd.extend([
+            '-V', 'monofont=DejaVuSansMono.ttf',
+            '-V', 'monofontoptions=BoldFont=DejaVuSansMono-Bold.ttf, ItalicFont=DejaVuSansMono-Oblique.ttf, BoldItalicFont=DejaVuSansMono-BoldOblique.ttf',
+        ])
     
     try:
         subprocess.run(cmd, check=True, capture_output=True)

package/package.json

@@ -1,10 +1,10 @@
 {
   "name": "@joaodotwork/md-2-pdf",
-  "version": "1.3.0",
+  "version": "1.4.0",
   "description": "Convert markdown files to PDF with mermaid diagram support",
   "main": "index.js",
   "bin": {
-    "md-2-pdf": "./bin/index.js"
+    "md-2-pdf": "bin/index.js"
   },
   "scripts": {
     "test": "echo \"Error: no test specified\" && exit 1"