gtparse fails on this file

[TLE]

When trying to use the latest trunk version (revision 65) of gtgrep on the attached file it fails

knielsen@knielsen-ThinkPad-T61:~/Dropbox/translations/evolution/3-2$ gtgrep -i reminder evolution.master.da.po
Traceback (most recent call last):
  File "/home/knielsen/code/pyg3t/bin/gtgrep", line 5, in <module>
    gtgrep.main()
  File "/home/knielsen/code/pyg3t/pyg3t/gtgrep.py", line 229, in main
    cat = parse(input)
  File "/home/knielsen/code/pyg3t/pyg3t/gtparse.py", line 553, in parse
    for chunk in chunk_iter(input, include_obsoletes=True):
  File "/home/knielsen/code/pyg3t/pyg3t/gtparse.py", line 503, in get_message_chunks
    line, msgctxt = _extract_string(line, input, patterns['msgctxt'])
  File "/home/knielsen/code/pyg3t/pyg3t/gtparse.py", line 467, in _extract_string
    string = extract_string(lines, header)
  File "/home/knielsen/code/pyg3t/pyg3t/gtparse.py", line 424, in extract_string
    raise PoSyntaxError('%s not found. Line was: %s' % (header, line))
pyg3t.gtparse.PoSyntaxError: msgctxt " not found. Line was: msgctxt "cal-reminders"

[TLE (k-nielsen81)]

Ok, so this is actually a gtparse error. The problem as I could see it is that the method extract_string does this:
if not line.startswith(header + ' "'):
but for some reason the header for the msgctxt already contains its ' "', so it can be fixed by this diff:

=== modified file 'pyg3t/gtparse.py'
--- pyg3t/gtparse.py 2011-06-19 17:20:03 +0000
+++ pyg3t/gtparse.py 2011-09-14 12:37:27 +0000
@@ -430,7 +430,7 @@
     return ''.join([headerline] + otherlines)

 linepatterns = dict(comment='#',
- msgctxt='msgctxt "',
+ msgctxt='msgctxt', # Used to be 'msgctxt "'
                     msgid='msgid',
                     msgid_plural='msgid_plural',
                     msgstr='msgstr',

However, I cannot really see if there is a particular reason that it was like that.

[TLE (k-nielsen81)]

After having fixed the bug above I ran into another one, which as far as I can see is unrelated:

Traceback (most recent call last):
  File "/home/knielsen/code/pyg3t/bin/podiff", line 5, in <module>
    podiff.main()
  File "/home/knielsen/code/pyg3t/pyg3t/podiff.py", line 278, in main
    cat_old = parse(open(args[0]))
  File "/home/knielsen/code/pyg3t/pyg3t/gtparse.py", line 554, in parse
    for chunk in chunk_iter(input, include_obsoletes=True):
  File "/home/knielsen/code/pyg3t/pyg3t/gtparse.py", line 503, in get_message_chunks
    if line.startswith(patterns['msgctxt']):
AttributeError: 'NoneType' object has no attribute 'startswith'

As far as I can tell it is triggerede when the parser encounters a translator created comment in the obsolete part. The part the triggers the fist error is this one:

# Jeg har fjernet mellemrum før ?, idet jeg antager at det er en fejl.
# Følg op på fejlrapport: https://bugzilla.gnome.org/show_bug.cgi?id=595002
#~ msgid ""
#~ "Retracting a message may remove it from the recipient's mailbox. Are you "
#~ "sure you want to do this?"
#~ msgstr ""
#~ "At tilbagetrække en besked kan fjerne den fra modtagerens brevbakke. Er "
#~ "du sikker på, at du ønsker at gøre dette?"

It also responds to fuzzys in the obsolotes, like this one:

#, fuzzy
#~ msgid "Migrate local mbox folders to maildir"
#~ msgstr "Opret en ny mappe til at gemme post i"

So I'm assuming that the problem is that there is content in the obsolete section that has no #~ in front of it. The problem is what to do with it. As far as I could tell from the gettext documentation page, the only requirement for an obsolete section is that the whole thing is a comment, and since that is true here, it may in fact be legal syntax.

However I'm unsure of how to fix this, or if having