[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [HEADS UP] Changes in date formatting (strftime, nl_langinfo)



28.01.2018 01:25 Ihar Hrachyshka <ihar.hrachyshka@gmail.com> wrote:
>
>
> On Sat, Jan 27, 2018 at 3:58 PM, Rafal Luzynski
> <digitalfreak@lingonborough.com> wrote:
> > [...]
> > OK, here I was asking if "Žnivień" -> "žniŭnia" is correct but now I can see
> > the same in Cyrillic in CLDR: "жнівень" -> "жніўня" which confirms that this
> > is correct.
>
> Yes, it's correct.

Nice.

> The upper/lower case letter depends on where the
> word is in the sentence. If it's the first word, it's upper case,
> otherwise lower (it's of course obvious, but it's better to be
> explicit). I don't know if Žnivień is always the first word, or not.

This is a different issue and it is confusing, indeed. Actually we need
a separate flag controlling whether the month name should be upper or
lower case. The nominative month names are sometimes uppercased because
they are expected to appear standalone but of course there are exceptions.

>
> >
> >> > vierasnia
> >>
> >> ^ there is a mistake here, it's vieraśnia, not vierasnia. Other names are
> >> fine.
> >
> > OK, I'm fixing locally and in github:
> >
> > https://github.com/rluzynski/glibc/commits/master
> >
> > Please review and... it looks that we can prepare the patch and push
> > it on Monday? Seems like we are going to fix the largest (by the number
> > of speakers) and the most affected languages.
>
> I think what should be fixed is:
>
> https://github.com/rluzynski/glibc/blob/master/localedata/locales/be_BY%40latin#L108

This says "vieraśnia". I guess you saw it before I fixed.

> Also this one is weird:
> https://github.com/rluzynski/glibc/blob/master/localedata/locales/be_BY%40latin#L159
> because if we apply Belarusian Latin script, it would be: biełaruskaja
> mova, but maybe I miss something.

OK, I'm going to fix this.

> As for Cyrillic versions, it's not easy to parse it, so I haven't
> checked.

It is imported from CLDR so most probably correct. If not please file a ticket
in CLDR. Also I'm going to post the decoded version online to simplify the
review.

> I actually think that the files should contain proper unicode
> strings that, if needed, would be converted back to U-notation for
> machine parsing, but maybe that's just me.

We were discussing this at glibc side and actually only recently
we have started using the actual ASCII characters because previously
everything consisted of the codes like <U0041>. Only ASCII because
there are so many scripts around the world that developers may be
unable to read, write, verify and maybe even display them.

Regards,

Rafal


Reply to: