fix: Handle missing date fields in XML submissions (#5744)

* refactor: Eliminate _construct_creation_date helper

* fix: Use xml2rfc method for filling in missing date fields

* fix: Set options.date for xml2rfc writers

* test: Test handling of missing date element/fields

ietf/submit/tests.py

@@ -3347,7 +3347,8 @@ def test_process_submission_xml(self):
             "test_submission.xml",
             title="Correct Draft Title",
         )
-        xml_path.write_text(xml.read())
+        xml_contents = xml.read()
+        xml_path.write_text(xml_contents)
         output = process_submission_xml("draft-somebody-test", "00")
         self.assertEqual(output["filename"], "draft-somebody-test")
         self.assertEqual(output["rev"], "00")
@@ -3362,6 +3363,23 @@ def test_process_submission_xml(self):
         self.assertIsNone(output["formal_languages"])
         self.assertEqual(output["xml_version"], "3")
 
+        # Should behave on missing or partial <date> elements
+        xml_path.write_text(re.sub(r"<date.+>", "", xml_contents))  # strip <date...> entirely
+        output = process_submission_xml("draft-somebody-test", "00")
+        self.assertEqual(output["document_date"], None)
+
+        xml_path.write_text(re.sub(r"<date year=.+ month", "<date month", xml_contents))  # remove year
+        output = process_submission_xml("draft-somebody-test", "00")
+        self.assertEqual(output["document_date"], date_today())
+
+        xml_path.write_text(re.sub(r"(<date.+) month=.+day=(.+>)", r"\1 day=\2", xml_contents))  # remove month
+        output = process_submission_xml("draft-somebody-test", "00")
+        self.assertEqual(output["document_date"], date_today())
+
+        xml_path.write_text(re.sub(r"<date(.+) day=.+>", r"<date\1>", xml_contents))  # remove day
+        output = process_submission_xml("draft-somebody-test", "00")
+        self.assertEqual(output["document_date"], date_today())
+
         # name mismatch
         xml, _ = submission_file(
             "draft-somebody-wrong-name-00",  # name that appears in the file

ietf/submit/utils.py

@@ -941,8 +941,10 @@ def render_missing_formats(submission):
         xmltree.tree = v2v3.convert2to3()
 
     # --- Prep the xml ---
+    today = date_today()
     prep = xml2rfc.PrepToolWriter(xmltree, quiet=True, liberal=True, keep_pis=[xml2rfc.V3_PI_TARGET])
     prep.options.accept_prepped = True
+    prep.options.date = today
     xmltree.tree = prep.prep()
     if xmltree.tree == None:
         raise SubmissionError(f'Error from xml2rfc (prep): {prep.errors}')
@@ -952,6 +954,7 @@ def render_missing_formats(submission):
     if not txt_path.exists():
         writer = xml2rfc.TextWriter(xmltree, quiet=True)
         writer.options.accept_prepped = True
+        writer.options.date = today
         writer.write(txt_path)
         log.log(
             'In %s: xml2rfc %s generated %s from %s (version %s)' % (
@@ -966,6 +969,7 @@ def render_missing_formats(submission):
     # --- Convert to html ---
     html_path = staging_path(submission.name, submission.rev, '.html')
     writer = xml2rfc.HtmlWriter(xmltree, quiet=True)
+    writer.options.date = today
     writer.write(str(html_path))
     log.log(
         'In %s: xml2rfc %s generated %s from %s (version %s)' % (

ietf/utils/draft.py

@@ -190,46 +190,6 @@ def get_title(self):
     def get_wordcount(self):
         raise NotImplementedError
 
-    @staticmethod
-    def _construct_creation_date(year, month, day=None):
-        """Construct a date for the document
-        
-        Roughly follows RFC 7991 section 2.17, but only allows missing day and
-        assumes the 15th if day is not specified month/year are not current.
-        
-        year: integer or string with 4-digit year
-        month: integer or string with numeric or English month. Some abbreviations recognized.
-        day: integer or string with numeric day of month. Optional.
-        
-        Raises ValueError if there is a problem interpreting the data
-        """
-        year = int(year)
-        day = int(day)
-        if isinstance(month, str):
-            month = month.lower()
-            if month in month_names:
-                month = month_names.index(month) + 1
-            elif month in month_names_abbrev3:
-                month = month_names_abbrev3.index(month) + 1
-            elif month in month_names_abbrev4:
-                month = month_names_abbrev4.index(month) + 1
-            elif month.isdigit() and int(month) in range(1, 13):
-                month = int(month)
-            else:
-                raise ValueError("Unrecognized month")
-        today = date_today()
-        if not day:
-            # if the date was given with only month and year, use
-            # today's date if month and year is today's month and
-            # year, otherwise pick the middle of the month.
-            # Don't use today's day for month and year in the past
-            if month == today.month and year == today.year:
-                day = today.day
-            else:
-                day = 15
-        return datetime.date(year, month, day)
-
-
 # ----------------------------------------------------------------------
 
 class PlaintextDraft(Draft):
@@ -500,7 +460,27 @@ def get_creation_date(self):
                 day = int( md.get( 'day', 0 ) )
                 year = int( md['year'] )
                 try:
-                    self._creation_date = self._construct_creation_date(year, mon, day)
+                    if   mon in month_names:
+                        month = month_names.index( mon ) + 1
+                    elif mon in month_names_abbrev3:
+                        month = month_names_abbrev3.index( mon ) + 1
+                    elif mon in month_names_abbrev4:
+                        month = month_names_abbrev4.index( mon ) + 1
+                    elif mon.isdigit() and int(mon) in range(1,13):
+                        month = int(mon)
+                    else:
+                        continue
+                    today = date_today()
+                    if day==0:
+                        # if the date was given with only month and year, use
+                        # today's date if month and year is today's month and
+                        # year, otherwise pick the middle of the month.
+                        # Don't use today's day for month and year in the past
+                        if month==today.month and year==today.year:
+                            day = today.day
+                        else:
+                            day = 15
+                    self._creation_date = datetime.date(year, month, day)
                     return self._creation_date
                 except ValueError:
                     # mon abbreviation not in _MONTH_NAMES

ietf/utils/xmldraft.py

@@ -1,12 +1,15 @@
 # Copyright The IETF Trust 2022, All Rights Reserved
 # -*- coding: utf-8 -*-
+import datetime
 import io
 import re
 import xml2rfc
 
 import debug  # pyflakes: ignore
 
 from contextlib import ExitStack
+from xml2rfc.util.date import augment_date, extract_date
+from ietf.utils.timezone import date_today
 
 from .draft import Draft
 
@@ -136,12 +139,19 @@ def get_title(self):
     def get_creation_date(self):
         date_elt = self.xmlroot.find("front/date")
         if date_elt is not None:
-            try:
-                year = date_elt.get("year")
-                month = date_elt.get("month")
-                return self._construct_creation_date(year, month, date_elt.get("day", None))
-            except ValueError:
-                pass
+            # ths mimics handling of date elements in the xml2rfc text/html writers
+            today = date_today()
+            year, month, day = extract_date(date_elt, today)
+            year, month, day = augment_date(year, month, day, today)
+            if day is None:
+                # Must choose a day for a datetime.date. Per RFC 7991 sect 2.17, we use
+                # today's date if it is consistent with the rest of the date. Otherwise,
+                # arbitrariy (and consistent with the text parser) assume the 15th.
+                if year == today.year and month == today.month:
+                    day = today.day
+                else:
+                    day = 15
+            return datetime.date(year, month, day)
         return None
 
     # todo fix the implementation of XMLDraft.get_abstract()