Improve rendering of minimal Markdown for description in package summary

Retain the heuristic which makes URLs in the description, and improve
it's handling of some cases.

4 files changed, 19 insertions, 11 deletions

diff --git a/calm/pkg2html.py b/calm/pkg2html.py
index d44b40b..ea4cb8f 100755
--- a/calm/pkg2html.py
+++ b/calm/pkg2html.py
@@ -52,6 +52,8 @@ import textwrap
 import time
 from typing import NamedTuple
 
+import markdown
+
 import xtarfile
 
 from . import common_constants
@@ -80,13 +82,18 @@ def ldesc(po, bv):
         return sdesc(po, bv)
 
     header = header.strip('"')
-    # escape html entities
-    header = html.escape(header, quote=False)
-    header = header.replace('\n\n', '\n<br>\n')
-    # try to recognize things which look like bullet points
-    header = re.sub(r'\n(\s*[*-]\s)', r'<br>\n\1', header)
-    # linkify things which look like hyperlinks
-    header = re.sub(r'http(s|)://[^\s\)]*', r'<a href="\g<0>">\g<0></a>', header)
+    header = markdown.markdown(header)
+
+    # linkify things which look like URLs
+    def linkify_without_fullstop(m):
+        url = m.group(0)
+        suffix = ''
+        if url[-1] == '.':
+            suffix = url[-1]
+            url = url[0:-1]
+        return '<a href="{0}">{0}</a>{1}'.format(url, suffix)
+
+    header = re.sub(r'http(s|)://[\w./_-]*', linkify_without_fullstop, header)
 
     return header
 
diff --git a/requirements.txt b/requirements.txt
index b75566e..3cb6f87 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -5,6 +5,7 @@ flake8-builtins
 flake8-import-order
 license_expression
 lockfile
+markdown
 pycodestyle
 python-daemon
 xtarfile[zstd]
diff --git a/test/testdata/htdocs.expected/summary/openssh-src.html b/test/testdata/htdocs.expected/summary/openssh-src.html
index 2bf497c..9ac2ab1 100755
--- a/test/testdata/htdocs.expected/summary/openssh-src.html
+++ b/test/testdata/htdocs.expected/summary/openssh-src.html
@@ -13,8 +13,8 @@
 <table class="pkgdetails">
 <tr><td><p><span class="detail">summary</span>:</p></td><td><p>The OpenSSH server and client programs</p></td></tr>
 <tr><td><p><span class="detail">description</span>:</p></td><td><p>OpenSSH is a program for logging into a remote machine and for
-	executing commands on a remote machine.  It can replace rlogin and rsh,
-	providing encrypted communication between two machines.</p></td></tr>
+    executing commands on a remote machine.  It can replace rlogin and rsh,
+    providing encrypted communication between two machines.</p></td></tr>
 <tr><td><p><span class="detail">categories</span>:</p></td><td><p>Net</p></td></tr>
 <tr><td><p><span class="detail">install package(s)</span>:</p></td><td><p><a href="openssh.html">openssh</a></p></td></tr>
 <tr><td><p><span class="detail">maintainer(s)</span>:</p></td><td><p>Corinna Vinschen
diff --git a/test/testdata/htdocs.expected/summary/openssh.html b/test/testdata/htdocs.expected/summary/openssh.html
index a97c3ca..75f3bc2 100644
--- a/test/testdata/htdocs.expected/summary/openssh.html
+++ b/test/testdata/htdocs.expected/summary/openssh.html
@@ -13,8 +13,8 @@
 <table class="pkgdetails">
 <tr><td><p><span class="detail">summary</span>:</p></td><td><p>The OpenSSH server and client programs</p></td></tr>
 <tr><td><p><span class="detail">description</span>:</p></td><td><p>OpenSSH is a program for logging into a remote machine and for
-	executing commands on a remote machine.  It can replace rlogin and rsh,
-	providing encrypted communication between two machines.</p></td></tr>
+    executing commands on a remote machine.  It can replace rlogin and rsh,
+    providing encrypted communication between two machines.</p></td></tr>
 <tr><td><p><span class="detail">categories</span>:</p></td><td><p>Net</p></td></tr>
 <tr><td><p><span class="detail">rdepends</span>:</p></td><td><p><a href="keychain.html">keychain</a></p></td></tr>
 <tr><td><p><span class="detail">source package</span>:</p></td><td><p><a href="openssh-src.html">openssh</a></p></td></tr>