[Bug localedata/27174] New: CLDR v38 and en_GB date issues

[Bug localedata/27174] New: CLDR v38 and en_GB date issues

[sourceware at robrwo dot sent.com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

            Bug ID: 27174
           Summary: CLDR v38 and en_GB date issues
           Product: glibc
           Version: unspecified
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: localedata
          Assignee: unassigned at sourceware dot org
          Reporter: sourceware at robrwo dot sent.com
                CC: libc-locales at sourceware dot org
  Target Milestone: ---

I ran across an issue with CLDR v38 that may affect the 
unicode-cldr-core package. (I believe the latest package used is v32 but 
it's worth being aware of this issue.)

One of the changes is for en_GB to change the date abbreviation for 
"September" from "Sep" to "Sept".  This will break software that is 
relying on locale information to parse or generate dates, and may also 
break any formatting that expects 3-letter months.

A thread which discusses how it affects a DateTime::Local perl module is 
at https://groups.perlists.pm/sympa/arc/london.pm/2021-01/msg00007.html 
but I imagine this will affect anything that relies on CLDR data.

I have reported the issue to the Unicode group at 
https://unicode-org.atlassian.net/browse/CLDR-14412

I have also noted the bug on the Ubuntu Translators list at
https://lists.ubuntu.com/archives/ubuntu-translators/2021-January/007729.html
where I was referred to here.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[sourceware at robrwo dot sent.com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #1 from Roibert Rothenberg <sourceware at robrwo dot sent.com> ---
Apologies. The text in the bug report was cut-and-paste from the ubuntu
translators list. It affects glibc.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
              Flags|                            |security-
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-01-12
             Status|UNCONFIRMED                 |WAITING
                 CC|                            |fweimer at redhat dot com

--- Comment #2 from Florian Weimer <fweimer at redhat dot com> ---
As far as I can see, we do not use “Sept“ in any of the en_* locales in glibc,
and never have.

Why do you think this is a glibc bug?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[sourceware at robrwo dot sent.com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #3 from Roibert Rothenberg <sourceware at robrwo dot sent.com> ---
(In reply to Florian Weimer from comment #2)
> As far as I can see, we do not use “Sept“ in any of the en_* locales in
> glibc, and never have.
> 
> Why do you think this is a glibc bug?

This was a recommendation from the ubuntu translators list.

Isn't strftime part of glibc?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[gunnarhj at ubuntu dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #4 from Gunnar Hjalmarsson <gunnarhj at ubuntu dot com> ---
Florian,

I was the one who suggested to bring this the glibc maintainers' attention. A
while ago the localedata in glibc were adapted to CLDR. Possibly that is
history.

But in any case, this kind of dissonance between CLDR/Unicode and locale data
doesn't feel healthy and may cause issues going forward.

Maybe it's not exactly a glibc bug at this time...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #5 from Florian Weimer <fweimer at redhat dot com> ---
A function called strftime is part of glibc, but the strftime used in the
report doesn't seem to call it:

https://metacpan.org/pod/DateTime#$dt-%3Estrftime(-$format,-...-)
https://github.com/houseabsolute/DateTime.pm/blob/d67ef93e1a5c31ec2fa1d0d4a06ca86b5073a048/lib/DateTime.pm#L1178

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #6 from Florian Weimer <fweimer at redhat dot com> ---
I do not see Sept in CLDR: https://st.unicode.org/cldr-apps/v#/en_GB/Gregorian/

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[sourceware at robrwo dot sent.com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #7 from Roibert Rothenberg <sourceware at robrwo dot sent.com> ---
I was posting this to point out that CLDR v38 has this issue, if/when strftime
is updated to use that data,

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[sourceware at robrwo dot sent.com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #8 from Roibert Rothenberg <sourceware at robrwo dot sent.com> ---
See https://github.com/unicode-org/cldr/blob/release-38/common/main/en_GB.xml

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #9 from Florian Weimer <fweimer at redhat dot com> ---
Ah, so this is basically a heads-up that CLDR 38 has some bogus data?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[sourceware at robrwo dot sent.com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #10 from Roibert Rothenberg <sourceware at robrwo dot sent.com> ---
(In reply to Florian Weimer from comment #9)
> Ah, so this is basically a heads-up that CLDR 38 has some bogus data?

Yes.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |NOTABUG
             Status|WAITING                     |RESOLVED

--- Comment #11 from Florian Weimer <fweimer at redhat dot com> ---
Thanks, I think we might spot this during review as part of the input even if
we forget about this bug.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug localedata/27174] CLDR v38 and en_GB date issues

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27174

--- Comment #12 from Florian Weimer <fweimer at redhat dot com> ---
I have changed my mind and do not consider switch to "Sept" as preposterous as
I initially indicated.

-- 
You are receiving this mail because:
You are on the CC list for the bug.