Multiset-ballot draftnames separated by comma only in the database

[ietf-svn-bot]

owner:olau@iola.dk resolution_fixed type_defect | by henrik@levkowetz.com

---

It seems that the multiset-ballot draft names are not always separated by
", ", as the code seems to expect, but instead in some cases by just ",".
The patch below seems to work in order to handle this better.

An example of a multi-ballot document is /doc/rfc4515/.

Index: ietf/idrfc/views_doc.py
===================================================================
--- ietf/idrfc/views_doc.py	(revision 3975)
+++ ietf/idrfc/views_doc.py	(working copy)
@@ -163,7 +163,7 @@
 
             multiset_ballot_text = "This was part of a ballot set with: "
             if e.desc.startswith(multiset_ballot_text):
-                names = e.desc[len(multiset_ballot_text):].split(", ")
+                names = [ n.strip() for n in e.desc[len(multiset_ballot_text):].split(",") ]
                 e.desc = multiset_ballot_text + ", ".join(u'<a href="%s">%s</a>' % (urlreverse("doc_view", kwargs={'name': n }), n) for n in names)
                 info["dontmolest"] = True

---

Issue migrated from trac:767 at 2022-03-04 01:58:15 +0000

[ietf-svn-bot]

@olau@iola.dk commented

---

This is some relatively old code I reviewed lately and decided it would be nicer with a space in between the draft names. But it looks like I've lost the accompanying patch for the importer somewhere so they are all imported without spaces (I was working on the charter branch when I found this, so I've probably managed to patch the importer in the wrong branch, then tested with that).

In any case, after some thought, I think the right approach is to fix the data by putting those spaces in. If you execute this snippet and revert the code, we should be fine:

from ietf.doc.models import DocEvent
for e in DocEvent.objects.filter(desc__regex="This was part of a ballot set with: .*,.*"):
 e.desc = "This was part of a ballot set with: " + ", ".join(e.desc[len("This was part of a ballot set with: "):].split(","))
 e.save()

I put in the spaces because I thought it was silly with that custom parsing hack, we should rather make a generic filter that just auto-links draft names.

[ietf-svn-bot]

@henrik@levkowetz.com commented

---

I agree on the generic filter, but I'm not sure I see why the generic filter shouldn't be written to handle any comma-separated list with arbitrary whitespace adjacent to the
commas?

[ietf-svn-bot]

@olau@iola.dk commented

---

Sorry, I'm getting ahead of myself. :)

I imagine a simple generic filter that matches a string like draft-foo-bar and turn it into a link like the urlize filter in Django. If we do that, we don't have to think about commas, but then this generated comment will look wrong because it says "draft-foo,draft-bar" instead of "draft-foo, draft-bar". So if we don't fix the missing space in there, we won't be able to make the filter so simple. Does that make sense?

I should perhaps clarify that it is the importer that has generated all of these comments to compensate for the loss of multi-ballot support in the new schema (so it's not a carry-over from the old data). I think I left out the space because I wasn't aware str.split can split on a whole string like ", ", and I hadn't thought of the generic filter at that time.

[ietf-svn-bot]

@henrik@levkowetz.com commented

---

Replying to ietf-svn-conversion/datatracker#767 (comment:3):

Sorry, I'm getting ahead of myself. :)

I imagine a simple generic filter that matches a string like draft-foo-bar and turn it into a link like the urlize filter in Django. If we do that, we don't have to think about commas, but then this generated comment will look wrong because it says "draft-foo,draft-bar" instead of "draft-foo, draft-bar". So if we don't fix the missing space in there, we won't be able to make the filter so simple. Does that make sense?

Ah. I understand. Something like what's mentioned in the first bullet here:
http://trac.tools.ietf.org/tools/ietfdb/wiki/IetfSpecificFeatures
:-)

I should perhaps clarify that it is the importer that has generated all of these comments to compensate for the loss of multi-ballot support in the new schema (so it's not a carry-over from the old data). I think I left out the space because I wasn't aware str.split can split on a whole string like ", ", and I hadn't thought of the generic filter at that time.

Ah. Understood. I'll apply the snippet and keep the code fix.

[ietf-svn-bot]

@henrik@levkowetz.com commented

---

Snipped applied. The deployed code still has this change, which is a bit more robust than the
original:

Index: ietf/idrfc/views_doc.py

--- ietf/idrfc/views_doc.py (revision 3997)
+++ ietf/idrfc/views_doc.py (working copy)
@@ -163,7 +163,7 @@

         multiset_ballot_text = "This was part of a ballot set with: "
         if e.desc.startswith(multiset_ballot_text):

•            names = e.desc[len(multiset_ballot_text):].split(", ")
  

•            names = [ n.strip() for n in e.desc[len(multiset_ballot_text):].split(",") ]
             e.desc = multiset_ballot_text + ", ".join(u'<a href="%s">%s</a>' % (urlreverse("doc_view", kwargs={'name': n }), n) for n in names)
             info["dontmolest"] = True
  

[ietf-svn-bot]

@henrik@levkowetz.com changed status from new to closed

[ietf-svn-bot]

@henrik@levkowetz.com changed resolution from `` to fixed

[ietf-svn-bot]

@henrik@levkowetz.com changed status from closed to deployed

[ietf-svn-bot]

@henrik@levkowetz.com commented

---

From e6c42e7:

Guard against trying to use .is_authenticated() on a plain string. The error condition can occur when an exception has been triggered before the auth middleware has run. Related to #767 and #768.

[ietf-svn-bot]

@henrik@levkowetz.com commented

---

From 8c9404e:

Guard against trying to use .is_authenticated() on a plain string. The error condition can occur when an exception has been triggered before the auth middleware has run. Related to #767 and #768.

[ietf-svn-bot]

@henrik@levkowetz.com commented

---

From 0011106:

Guard against trying to use .is_authenticated() on a plain string. The error condition can occur when an exception has been triggered before the auth middleware has run. Related to #767 and #768.

[ietf-svn-bot]

@henrik@levkowetz.com commented

---

From 14e2af1:

Committing patch from #767, to make splitting of comma-separated multiset-ballot draft-names more robust. Related to #767