[asterisk-dev] Asterisk 1.8 and character sets

[Örn Arnarson]

As far as I can tell, it seems to be perfectly valid to escape the
characters, and reserved and control characters in URIs MUST be
escaped according to the RFC. However, I don't understand why it's
done on the display name.

SIP is in UTF-8, and the escaped character is escaped as an UTF-8
character (Ö becomes %C3%96 instead of %D6), so if the idea is to make
it compatible with devices that don't understand UTF-8, it doesn't
really make sense. It won't know how to display an escaped UTF-8
character any better than it would if it were simply printed as it is.

There was a big discussion on these changes that I found here:
https://issues.asterisk.org/view.php?id=16299

As far as I can tell, the intention was _NOT_ to URI encode the
Display name, as oej pointed out:
"Again, the display name should *NOT* be URI-encoded. Ever. It's not
part of a URI."

This, however, seems to be the case, and it doesn't seem like anyone
addressed what he said prior to submitting the patch.

I'm not entirely sure what the correct method of doing is this
according to the RFC. In theory, escaping it should probably be fine
(albeit unnecessary), but I'm not certain since the Display name isn't
an URI, and is contained within "".

What I do know, however, is that Asterisk 1.8 breaks displaying the
caller-id name correctly on all Aastra terminals that I've tested in
on. It also breaks it on Doro phones. I will try to test Linksys and
Snom tomorrow, but I suspect it'll have the same results. I don't
think the Display name is meant to be URI escaped. At the very least,
it is completely unnecessary as SIP is already in UTF-8.

Regards,
Örn

2010/11/10 Klaus Darilion <klaus.mailinglists at pernau.at>:
>
>
> On 03.11.2010 11:02, Örn Arnarson wrote:
>>
>> I sent this first to the Asterisk users list, but this question
>> probably belongs on the dev list, so here goes:
>>
>> Caller-ID behavior is different -- now when I set the caller-id name
>> to something with special characters (Ö, for example), the SIP INVITE
>> now has %C3%96 instead of the Ö character. I've tried doing
>> Set(CALLERID(name-charset)=utf8) as well as iso8859-1, but it's always
>> the same behavior.
>>
>> Here's the header as it appears in 1.6.2.11 CLI output:
>>
>> INVITE sip:1502 at 192.168.10.169:5060;transport=udp SIP/2.0
>> Via: SIP/2.0/UDP 192.168.10.3:5060;branch=z9hG4bK73713002;rport
>> Max-Forwards: 70
>> From: "SIP ehf/Örn Arnarson"<sip:7712552 at 192.168.10.3>;tag=as2813a8fe
>> To:<sip:1502 at 192.168.10.169:5060;transport=udp>
>> Contact:<sip:7712552 at 192.168.10.3>
>>
>> Here is the header with the same caller-id information in 1.8.0:
>>
>> INVITE sip:1502 at 192.168.10.169:5060;transport=udp SIP/2.0
>> Via: SIP/2.0/UDP 192.168.10.3:5060;branch=z9hG4bK1ff411d7
>> Max-Forwards: 70
>> From: "SIP ehf/%C3%96rn Arnarson"<sip:7712552 at 192.168.10.3>;tag=as701c8835
>> To:<sip:1502 at 192.168.10.169:5060;transport=udp>
>> Contact:<sip:7712552 at 192.168.10.3:5060>
>
> Have you checked the RFC? I guess that UTF-8 is allowed in the display name,
> but %-encoding may be allowed too.
>
> regards
> klaus
>