Preprocess signature separately

[etc/mutt.git] / .mutt / markdown2html

diff --git a/.mutt/markdown2html b/.mutt/markdown2html
index 72d0830dcf02149917169280e43a8ccb0a9f519f..8930c88cc91d3bdb15383f1c709b69978b63a559 100755 (executable)
--- a/.mutt/markdown2html
+++ b/.mutt/markdown2html
@@ -116,6 +116,12 @@ SIGNATURE_HTML = \
         '<div class="signature"><span class="leader">-- </span>{sig}</div>'
 
 
+def _preprocess_signature(sig):
+    '''
+    Preprocess the signature before markdown processing.
+    '''
+    return sig
+
 def _preprocess_markdown(mdwn):
     '''
     Preprocess Markdown for handling by the converter.
@@ -133,6 +139,13 @@ def _preprocess_markdown(mdwn):
     # text-mode HTML2text converters, and so it's left commented for now.
     #ret = re.sub(r'\n>', r'  \n>[>]{.quotechar}', ret, flags=re.MULTILINE)
 
+    # With the autolink_bare_uris extension, we do not need to put links into
+    # angle brackets to have them converted, so let's conserve the brackets
+    # when used around email addresses. Note that this needs a postprocessing
+    # hack because the pandoc autolink converted includes the ambersand
+    # (https://github.com/jgm/pandoc/issues/7398).
+    ret = re.sub(r'<([^@]+@.+\.[^>]+)>', r'&lt;\g<1> -PANDOC_BUG_7398-&gt;', ret)
+
     return ret
 
 
@@ -214,7 +227,7 @@ def _reformat_quotes(html):
 
 def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
                          ext_enabled=None, ext_disabled=None,
-                         standalone=True, title="HTML E-Mail"):
+                         standalone=True, selfcontained=True, title=None):
     '''
     Invoke pandoc to do the actual conversion of Markdown to HTML5.
     '''
@@ -257,6 +270,8 @@ def _convert_with_pandoc(mdwn, inputfmt='markdown', outputfmt='html5',
     args = []
     if standalone:
         args.append('--standalone')
+    if selfcontained:
+        args.append('--self-contained')
     if title:
         args.append(f'--metadata=pagetitle:"{title}"')
 
@@ -275,6 +290,10 @@ def _postprocess_html(html):
     '''
     Postprocess the generated and styled HTML.
     '''
+
+    # Preprocessing leaves a sentinel to work around
+    # https://github.com/jgm/pandoc/issues/7398, and so we need to remove it:
+    html = html.replace('</a> -PANDOC_BUG_7398-&gt;', '</a>&gt;')
     return html
 
 
@@ -294,12 +313,16 @@ def convert_markdown_to_html(mdwn):
     if body:
         body = _preprocess_markdown(body)
         body = _identify_quotes_for_later(body)
-        html = _convert_with_pandoc(body, standalone=False)
+        html = _convert_with_pandoc(body, standalone=True, selfcontained=True)
         html = _reformat_quotes(html)
 
     if sig:
+        sig = _preprocess_signature(sig)
         sig = _preprocess_markdown(sig)
-        html += SIGNATURE_HTML.format(sig='<br/>'.join(sig.splitlines()))
+        sig = _convert_with_pandoc(sig, standalone=False, selfcontained=False)
+        sig = SIGNATURE_HTML.format(sig='<br/>'.join(sig.splitlines()))
+        eob = html.find('</body>')
+        html = f'{html[:eob]}{sig}\n{html[eob:]}'
 
     html = HTML_DOCUMENT.format(htmlbody=html)
     html = _apply_styling(html)