From c0f170462209484074470d693af44c65e5a252ac Mon Sep 17 00:00:00 2001
From: Skip Montanaro <skip.montanaro@gmail.com>
Date: Tue, 16 Feb 2021 20:17:05 -0600
Subject: simple script to extract (most of) the original shell archive from
 the Google Groups HTML

---
 shar/sharify.py | 55 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 55 insertions(+)
 create mode 100644 shar/sharify.py

(limited to 'shar')

diff --git a/shar/sharify.py b/shar/sharify.py
new file mode 100644
index 0000000..42ec1c4
--- /dev/null
+++ b/shar/sharify.py
@@ -0,0 +1,55 @@
+#!/usr/bin/env python
+
+"""A better reconstitution of the original shell archives from the Google HTML files.
+
+Example (for part 09):
+
+curl https://groups.google.com/g/alt.sources/c/w0LgGPVB6f0/m/SDnD377as9IJ \
+| egrep '<section>' \
+| sed -e 's/.*<section>//' -e 's:</section>.*::' \
+| python shar/sharify.py > shar/python-0.9.1-09-21.shar
+"""
+
+import html
+import re
+import sys
+
+shar = sys.stdin.read()
+
+for (pat, repl) in (
+        (r'X\t', r'X	'),
+        (r'X	\t', r'X		'),
+        (r'X		\t', r'X			'),
+        (r'X			\t', r'X				'),
+        ):
+    shar = shar.replace(pat, repl)
+
+for (pat, repl) in (
+        (',[2,[[1,[null,"', ''),
+        (r'\u0026lt;\u003c', '<'),
+        (r'\u003e\u0026gt;', '>'),
+        (r'\u003c', '<'),
+        (r'\u003d', '='),
+        (r'\u003e', '>'),
+        (r'\u0026', '&'),
+        (r'\"', '"'),
+        (r'"]', ''),
+):
+    shar = shar.replace(pat, repl)
+
+shar = html.unescape(shar)
+
+for (pat, repl) in (
+        ('<br>', '\n'),
+        ):
+    shar = shar.replace(pat, repl)
+
+for (pat, repl) in (
+        (r'<a href="[^"]+" target="_blank" rel="nofollow"'
+         r' data-saferedirecturl="[^"]+">([^<]+)</a>',
+         r'\1'),
+        (r'<a href data-email-masked rel="nofollow">([^<]+)</a>', r'\1'),
+    ):
+    shar = re.sub(pat, repl, shar)
+
+sys.stdout.write(shar.rstrip())
-- 
cgit v1.2.3