Re: [I18ndir] Review of new characters for Unicode 12.0.0

["Patrik Fältström " <[e-mail]>]

On 19 Mar 2019, at 0:47, Martin J. Dürst wrote:

>> Martin, as I have implemented IDNA2008 from scratch, i.e. not using any
>> libraries at all,
>
> I'm also not using any libraries at all, just relying on the Unicode version in Ruby to be the same as the target version. But I understand what you're saying.

Well, Ruby do use ultimately libidn2 for IDNA2008, so yes, you do use libraries.

>> let me suggest we sync on what output format we use so
>> that we can do "diff" between your and my lists. My list for 12.0.0 is
>> btw attached.
>
> I think it's a good idea to have this for some cross-checks. But one of the features of my program was that it listed only the new characters, because these were the ones I wanted to review. So I'll definitely keep this ability.

Ok, I of course have that as well :-)

That operates on the file format I did show you.

>> May I suggest what my program is doing (of course), which is as follows:
>>
>> <Codepoint>;<Derived property value>;<Rule(s) that matched>;<Name>
>
> I'm fine with that. I'll have to do more work on supporting ranges in UnicodeData.txt and supporting the rules that I ignored in my first quick attempt. So please don't rely on my work for moving forward.

No problem, I already do have two implementations myself. One that do use libidn2 in python3, one in ruby that do not, and then I compare with IANA as I know they do have a completely code base as well.

   Patrik