Re: [I18n-discuss] Is this an A-label or what?

[Martin J. Dürst <[e-mail]>]

On 2019/02/14 00:42, John R Levine wrote:
> This does what you'd expect, two A-labels turn into two U-labels:
> 
>>>> idn2.atou('xn--eckwd4c7c.xn--zckzah')
> 'ドメイン.テスト'
> 
> But someone added a space at the end.  Gnu idn2 decodes it as punycode, 
> presumably using some implicit negative code value for the space:
> 
>>>> idn2.atou('xn--eckwd4c7c.xn--zckzah ')
> 'ドメイン.テニスト'

Just for those who don't read Japanese, and don't trust their eyes with 
details in foreign scripts: The two results are clearly different. The 
first reads "domain.test", the second reads "domain.tenist" (or some 
such, obviously, "tenist" or "tenisuto" in a more literal transcription 
is not a word). The "ニ" is clearly superfluous.

> The existing python code barfs because it doesn't expect a space:
> 
> idna.decode('xn--eckwd4c7c.xn--zckzah ')
> IndexError: string index out of range

That's a good thing to do.

> The punycode algorithm will never create a label with a space at the 
> end, If you see one do you decode it anyway?  Treat it as some random 
> ASCII label and pass it through?  Reject it because it looks sort of 
> like an A-label but it's not a hostname?

I could expect a library in a some more lenient language to just ignore 
the space. Using the space as part of the calculation seems really 
weird. Of course Python, being a rather strict language, barfs.

Regards,   Martin.

> Regards,
> John Levine, [e-mail], Taughannock Networks, Trumansburg NY
> Please consider the environment before reading this e-mail. https://jl.ly