Opened 8 years ago

Closed 8 years ago

#91 closed defect (fixed)

Mapping of <ireg-name>: preferred way?

Reported by: evnikita2@… Owned by: duerst@…
Priority: critical Milestone:
Component: 3987bis Version:
Severity: - Keywords:
Cc: masinter@…

Description

From http://lists.w3.org/Archives/Public/public-iri/2011Aug/0039.html:


Section 3.4: I may be wrong, but why we mandate use of pct-encoding when
mapping <ireg-name> with SHOULD and IDNA procedure is optional? If
there are some reasons for preferring the former, please explain it in
the document.

Change History (7)

comment:1 Changed 8 years ago by masinter@…

  • Priority changed from minor to critical

I think this is a major issue that needs resolving. My belief (Larry) is that punicode mapping is a (strong) SHOULD, or even a MUST, and that the %xx encoding has too much risk of sending percent-encoded names to unaware DNS components.

In fact, this recommendation was the basis for changing the spec to parse IRI components first.

comment:2 Changed 8 years ago by duerst@…

I think that pct-encoding is the right thing in the long term. Some implementations are already doing it, others will follow. I personally don't want a spec that locks us into some special case forever when we are moving towards getting rid of it.

comment:3 Changed 8 years ago by masinter@…

  • Owner changed from draft-ietf-iri-3987bis@… to masinter@…

I propose resolving this by being more specific about exactly when punicode encoding is acceptable, by changing

OLD:

<t>The ireg-name component MAY also be converted as follows:</t>

NEW:

<t>In situations where it is certain that the ireg-name is intended to be used as a domain name to be processed by Domain Name Lookup (as per <xref target="RFC53891"/>, an alternative method MAY may be used for converting the ireg-name component as follows: </t>

Is this OK?

comment:4 Changed 8 years ago by masinter@…

Martin suggests we add an introductory sentence noting that there are two ways... just so that the section isn't confusing for those new to the issue. This sounds reasonable, but I hope Martin will propose wording.

comment:5 Changed 8 years ago by masinter@…

  • Cc masinter@… added
  • Owner changed from masinter@… to duerst@…

comment:6 Changed 8 years ago by duerst@…

Here is my proposal for the introductory sentence, which is very simple:

--- C:/Users/duerst/AppData/Local/Temp/draft-ietf-iri-3987bis.xml-revBASE.svn000.tmp.xml	木 3  1 19:10:21 2012
+++ C:/Data/ietf-iri/draft-ietf-iri-3987bis/draft-ietf-iri-3987bis.xml	木 3  1 19:51:01 2012
@@ -652,6 +652,8 @@
 </section> <!-- general conversion -->
 
 <section title="Mapping ireg-name" anchor="dnsmapping">
+  <t>The mapping from &lt;ireg-name> to a &lt;reg-name> requires a choice
+  between one of the two methods described below.</t>
   <section title='Mapping using Percent-Encoding' anchor='dnspercent'>
   <t>The ireg-name component SHOULD be converted
     according to the general procedure for percent-encoding

comment:7 Changed 8 years ago by duerst@…

  • Resolution set to fixed
  • Status changed from new to closed

addressed with commit 99

Note: See TracTickets for help on using tickets.