From 30c64e052c508f945d001e5b6020e4de69ec6441 Mon Sep 17 00:00:00 2001
From: Luke Dunstan <LukeDunstan81@gmail.com>
Date: Wed, 10 Jun 2015 23:04:45 +0800
Subject: [PATCH] Helper script unextract.py

---
 python/unextract.py | 135 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 135 insertions(+)
 create mode 100755 python/unextract.py

diff --git a/python/unextract.py b/python/unextract.py
new file mode 100755
index 0000000..846c169
--- /dev/null
+++ b/python/unextract.py
@@ -0,0 +1,135 @@
+#!/usr/bin/env python3
+
+import os
+import os.path
+import re
+import xml.etree.ElementTree as etree
+
+from rfc_notes import ParseException
+NS = ''
+
+
+REFTYPES = { 'impl' : 'Impl', 'test' : 'Test' }
+
+class Reference:
+    def __init__(self, linenum, id):
+        self.linenum = linenum
+        self.id = id
+
+
+class SourceFile:
+    def __init__(self, path, reftype):
+        self.reftype = reftype
+        self.refs = []
+
+    def insert_refs(self, lines, docref, offset):
+        for ref in self.refs:
+            i = ref.linenum - 1 + offset
+            attr = ' [@ref {0} "{1}"]'.format(docref, ref.id)
+            line = lines[i].rstrip('\n')
+            if line.find(attr) == -1:
+                lines[i] = line + attr + '\n'
+
+
+def get_target(elem, parent_map):
+    assert elem.tag == NS + 'coderef'
+    parent = parent_map[elem]
+    assert parent.tag == NS + 'notes'
+    grandparent = parent_map[parent]
+    assert grandparent.tag in (NS + 'clause', NS + 'paragraph', NS + 'section')
+    return grandparent.get('id')
+
+
+def add_header(lines, header):
+    offset = 0
+    i = 0
+    # Skip comments, blank lines and existing attributes at the beginning of the file
+    while i < len(lines):
+        line = lines[i].lstrip()
+        if line.startswith('(*'):
+            start_comment = i
+            # Find the end of the comment
+            while i < len(lines) and lines[i].find('*)') == -1:
+                # TODO: support nested comments
+                assert lines[i].find('(*') == -1 or i == start_comment
+                i += 1
+        elif line:
+            if line.find('[@') == -1:
+                break
+            elif line in header:
+                header.remove(line)
+        i += 1
+    if header:
+        for line in header:
+            lines.insert(i, line)
+            offset += 1
+            i += 1
+        lines.insert(i, '\n')
+        offset += 1
+    return offset
+
+
+def insert_attributes(xml, target_dir, let_name, camlp4):
+    if not let_name:
+        let_name = 'rfc'
+    parent_map = {child:parent for parent in xml.iter() for child in parent}
+    root = xml.getroot()
+    if root.tag != NS + 'rfc':
+        raise ParseException('Specification XML file should have <rfc> as root')
+    rfc_num = root.attrib['number']
+
+    files = {}
+    #import pdb; pdb.set_trace()
+    for elem in root.iter():
+        if elem.tag == NS + 'coderef':
+            path = elem.get('path')
+            reftype = REFTYPES[elem.get('type')]
+            if path not in files:
+                file = SourceFile(path, reftype)
+                files[path] = file
+            else:
+                file = files[path]
+                assert file.reftype == reftype
+            file.refs.append(Reference(int(elem.get('line')), get_target(elem, parent_map)))
+
+    for rel_path, file in files.items():
+        print(rel_path)
+        path = os.path.join(target_dir, rel_path)
+        with open(path, 'r') as f:
+            lines = f.readlines()
+        if camlp4:
+            offset = 0
+            # camlp4 mangles [@ref (rfc 9999) "xxx"] into [@ref rfc 999 "xxx"]
+            #docref = '(rfc {0})'.format(rfc_num)
+            docref = let_name
+        else:
+            header = [
+                    '[@@@reftype {0}]\n'.format(file.reftype),
+                    '[@@@specdoc let {0} = rfc {1}"]\n'.format(let_name, rfc_num),
+                    ]
+            offset = add_header(lines, header)
+            docref = let_name
+        file.insert_refs(lines, docref, offset)
+        with open(path, 'w') as f:
+            f.writelines(lines)
+
+
+def main():
+    import argparse
+    parser = argparse.ArgumentParser(description='Read manually edited code references from an annotated IETF RFC in XML and add them to OCaml code as attributes')
+    parser.add_argument('input', metavar='rfcNNNN_notes.xml', nargs=1, type=str,
+            help='The path to the input XML document (.xml)')
+    parser.add_argument('--let', dest='let', type=str,
+            help='The name to bind with [@@@specdoc let name = rfc <number>]')
+    parser.add_argument('--target', dest='target', type=str,
+            help='The root directory containing the source code to be modified')
+    parser.add_argument('--camlp4', action='store_true',
+            help='Set this option to generate camlp4-compatible code')
+    args = parser.parse_args()
+
+    doc = etree.parse(args.input[0])
+
+    insert_attributes(doc, args.target, args.let, args.camlp4)
+
+if __name__ == '__main__':
+    main()