public inbox for libc-locales@sourceware.org
 help / color / mirror / Atom feed
* Incorrect days of the week in ru_UA locale
@ 2019-09-27 15:25 Mikhail Gribanov
  2019-09-27 15:44 ` Florian Weimer
  2019-09-27 22:49 ` Rafal Luzynski
  0 siblings, 2 replies; 14+ messages in thread
From: Mikhail Gribanov @ 2019-09-27 15:25 UTC (permalink / raw)
  To: libc-alpha, bug-glibc-locales

Hi, sorry if I wrote the wrong way, I did not find who to contact to
fix this problem.

Preamble: Many linux distributions use the wrong day of the week in
ru_UA locale and close my bug reports because of "out of scope". After
some searching I managed to find out that localization is taken from
glibc. I hope you can help me (or send to the right place).

Problem: In ru_UA locale for days of the week the abbreviation with
three letters is used, which is incorrect: "Пнд", "Вто", "Срд"...

Solution/Fix: Use two-letter abbreviations:"Пн", "Вт", "Ср", "Чт",
"Пт", "Сб", "Вс". You can also just copy days of the week from ru_RU
locale where are they represented correctly.

Thank you for your time.

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-09-27 15:25 Incorrect days of the week in ru_UA locale Mikhail Gribanov
@ 2019-09-27 15:44 ` Florian Weimer
  2019-09-27 22:49 ` Rafal Luzynski
  1 sibling, 0 replies; 14+ messages in thread
From: Florian Weimer @ 2019-09-27 15:44 UTC (permalink / raw)
  To: Mikhail Gribanov; +Cc: libc-alpha, bug-glibc-locales

* Mikhail Gribanov:

> Hi, sorry if I wrote the wrong way, I did not find who to contact to
> fix this problem.
>
> Preamble: Many linux distributions use the wrong day of the week in
> ru_UA locale and close my bug reports because of "out of scope". After
> some searching I managed to find out that localization is taken from
> glibc. I hope you can help me (or send to the right place).
>
> Problem: In ru_UA locale for days of the week the abbreviation with
> three letters is used, which is incorrect: "Пнд", "Вто", "Срд"...
>
> Solution/Fix: Use two-letter abbreviations:"Пн", "Вт", "Ср", "Чт",
> "Пт", "Сб", "Вс". You can also just copy days of the week from ru_RU
> locale where are they represented correctly.

Does the choice have political implications?  uk_UA locale uses
three-letter month abbreviations, I think.

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-09-27 15:25 Incorrect days of the week in ru_UA locale Mikhail Gribanov
  2019-09-27 15:44 ` Florian Weimer
@ 2019-09-27 22:49 ` Rafal Luzynski
  2019-09-28 17:25   ` Keld Simonsen
  1 sibling, 1 reply; 14+ messages in thread
From: Rafal Luzynski @ 2019-09-27 22:49 UTC (permalink / raw)
  To: Mikhail Gribanov, libc-alpha, bug-glibc-locales

27.09.2019 17:24 Mikhail Gribanov <mikhail.gribanov@gmail.com> wrote:
> 
> Hi, sorry if I wrote the wrong way, I did not find who to contact to
> fix this problem.

Bugzilla [1] seems to be the best place but here it is equally good.
After all, your voice is being heard.

> Preamble: Many linux distributions use the wrong day of the week in
> ru_UA locale and close my bug reports because of "out of scope". After
> some searching I managed to find out that localization is taken from
> glibc. I hope you can help me (or send to the right place).

Correct.

> Problem: In ru_UA locale for days of the week the abbreviation with
> three letters is used, which is incorrect: "Пнд", "Вто", "Срд"...
> 
> Solution/Fix: Use two-letter abbreviations:"Пн", "Вт", "Ср", "Чт",
> "Пт", "Сб", "Вс". You can also just copy days of the week from ru_RU
> locale where are they represented correctly.
> 
> Thank you for your time.

Thank you for spotting a reporting the problem.

Indeed, ru_RU provides two-letter abbreviations as you quoted
while CLDR [2] provides the same but says they should start with
lowercase letters.  But the lowercase/uppercase issue is always
controversial in the languages where names of weekdays and months
do not start with an uppercase letter by default and no solution
is perfect so I assume that it's better to leave it uppercase,
as it is now.

When you said "copy days of the week from ru_RU" probably you meant
literally copy and paste, but given that there are almost no differences
in LC_TIME between ru_RU and ru_UA we can also use “copy "ru_RU"”
in LC_TIME in ru_UA.  Please note that it is impossible to copy part
of the section and impossible to modify it, either copy whole section
as it is or keep two separate copies.

There is only one other difference between ru_RU and ru_UK:
in ru_RU ab_alt_mon (abbreviated month names in nominative case)
begin with a lowercase letter while in ru_UK they begin with an
uppercase.  Is it OK to unify them?  Uppercase looks better for me
in this case, OK to make it uppercase, that is copy ab_alt_mon
from ru_UK to ru_RU?

If you are curious where to see ab_alt_mon in real applications
the answer is: in calendar headers (if they are abbreviated).
Never in full date formats.

Regards,

Rafal


[1] https://sourceware.org/bugzilla/enter_bug.cgi?product=glibc
[2] https://st.unicode.org/cldr-apps/v#/ru/Gregorian/

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-09-27 22:49 ` Rafal Luzynski
@ 2019-09-28 17:25   ` Keld Simonsen
  2019-10-01 19:45     ` Rafal Luzynski
  0 siblings, 1 reply; 14+ messages in thread
From: Keld Simonsen @ 2019-09-28 17:25 UTC (permalink / raw)
  To: Rafal Luzynski; +Cc: Mikhail Gribanov, libc-alpha, bug-glibc-locales

I think we should use lowercase that is what we do with manuyothe european locales
and it is the culturally correct form

keld

On Sat, Sep 28, 2019 at 12:48:20AM +0200, Rafal Luzynski wrote:
> 27.09.2019 17:24 Mikhail Gribanov <mikhail.gribanov@gmail.com> wrote:
> > 
> > Hi, sorry if I wrote the wrong way, I did not find who to contact to
> > fix this problem.
> 
> Bugzilla [1] seems to be the best place but here it is equally good.
> After all, your voice is being heard.
> 
> > Preamble: Many linux distributions use the wrong day of the week in
> > ru_UA locale and close my bug reports because of "out of scope". After
> > some searching I managed to find out that localization is taken from
> > glibc. I hope you can help me (or send to the right place).
> 
> Correct.
> 
> > Problem: In ru_UA locale for days of the week the abbreviation with
> > three letters is used, which is incorrect: "??????", "??????", "??????"...
> > 
> > Solution/Fix: Use two-letter abbreviations:"????", "????", "????", "????",
> > "????", "????", "????". You can also just copy days of the week from ru_RU
> > locale where are they represented correctly.
> > 
> > Thank you for your time.
> 
> Thank you for spotting a reporting the problem.
> 
> Indeed, ru_RU provides two-letter abbreviations as you quoted
> while CLDR [2] provides the same but says they should start with
> lowercase letters.  But the lowercase/uppercase issue is always
> controversial in the languages where names of weekdays and months
> do not start with an uppercase letter by default and no solution
> is perfect so I assume that it's better to leave it uppercase,
> as it is now.
> 
> When you said "copy days of the week from ru_RU" probably you meant
> literally copy and paste, but given that there are almost no differences
> in LC_TIME between ru_RU and ru_UA we can also use ???copy "ru_RU"???
> in LC_TIME in ru_UA.  Please note that it is impossible to copy part
> of the section and impossible to modify it, either copy whole section
> as it is or keep two separate copies.
> 
> There is only one other difference between ru_RU and ru_UK:
> in ru_RU ab_alt_mon (abbreviated month names in nominative case)
> begin with a lowercase letter while in ru_UK they begin with an
> uppercase.  Is it OK to unify them?  Uppercase looks better for me
> in this case, OK to make it uppercase, that is copy ab_alt_mon
> from ru_UK to ru_RU?
> 
> If you are curious where to see ab_alt_mon in real applications
> the answer is: in calendar headers (if they are abbreviated).
> Never in full date formats.
> 
> Regards,
> 
> Rafal
> 
> 
> [1] https://sourceware.org/bugzilla/enter_bug.cgi?product=glibc
> [2] https://st.unicode.org/cldr-apps/v#/ru/Gregorian/

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-09-28 17:25   ` Keld Simonsen
@ 2019-10-01 19:45     ` Rafal Luzynski
  2019-10-01 20:54       ` Keld Simonsen
  0 siblings, 1 reply; 14+ messages in thread
From: Rafal Luzynski @ 2019-10-01 19:45 UTC (permalink / raw)
  To: Keld Simonsen; +Cc: Mikhail Gribanov, libc-alpha, bug-glibc-locales

28.09.2019 19:37 Keld Simonsen <keld@keldix.com> wrote:
> 
> I think we should use lowercase that is what we do with manuyothe european
> locales
> and it is the culturally correct form
> 
> keld

Thank you for your feedback, Keld.  Unfortunately, as a native speaker
of one of those languages I must say that this is more complex.
Indeed, names of the weeks and months are usually lowercase but
when they appear in the beginning of a sentence they are uppercase.
I am afraid that a name of a weekday more often appears in the beginning
of the sentence than in the middle.  To the extent that some locales
prefer the weekday and/or month names to start with the uppercase.

What we actually need is a new format modifier to control the first
letter being uppercase/lowercase.  Currently we have only "^" to convert
whole substring (e.g., a weekday name) to uppercase, and "#" to swap
the upper/lowercase.  It was discussed in the past, there is a bug
report and a proposal to ensure that "^#" works as a converter to
lowercase but this is not enough because I would like to see one more
format modifier to convert to titlecase (that is: the first letter
being uppercase).  I would be happy to work on that problem but
it is a separate case and would need more discussion.

Back about the main problem, I would like to post a patch which would
be the best explanation what I mean in this particular case.

Best regards,

Rafal

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-01 19:45     ` Rafal Luzynski
@ 2019-10-01 20:54       ` Keld Simonsen
  2019-10-02  7:33         ` Mikhail Gribanov
  2019-10-02 22:14         ` Rafal Luzynski
  0 siblings, 2 replies; 14+ messages in thread
From: Keld Simonsen @ 2019-10-01 20:54 UTC (permalink / raw)
  To: Rafal Luzynski; +Cc: Mikhail Gribanov, libc-alpha, bug-glibc-locales

On Tue, Oct 01, 2019 at 09:45:13PM +0200, Rafal Luzynski wrote:
> 28.09.2019 19:37 Keld Simonsen <keld@keldix.com> wrote:
> > 
> > I think we should use lowercase that is what we do with manuyothe european
> > locales
> > and it is the culturally correct form
> > 
> > keld
> 
> Thank you for your feedback, Keld.  Unfortunately, as a native speaker
> of one of those languages I must say that this is more complex.
> Indeed, names of the weeks and months are usually lowercase but
> when they appear in the beginning of a sentence they are uppercase.
> I am afraid that a name of a weekday more often appears in the beginning
> of the sentence than in the middle.  To the extent that some locales
> prefer the weekday and/or month names to start with the uppercase.

I am also a native speaker of one of those languages, Danish, and moreover I was the initial contributer of about 60 of the locales incl ru_RU,
that Urich Drepper picked up for glibc. I am also the editor of ISO 14652 and ISO 30112 that provided extensions
beyond Posix i18n functionality, and I  provide guidance for their use. My advice is that this bad policy to have wrongly
capitalized names here. 

> What we actually need is a new format modifier to control the first
> letter being uppercase/lowercase.  Currently we have only "^" to convert
> whole substring (e.g., a weekday name) to uppercase, and "#" to swap
> the upper/lowercase.  It was discussed in the past, there is a bug
> report and a proposal to ensure that "^#" works as a converter to
> lowercase but this is not enough because I would like to see one more
> format modifier to convert to titlecase (that is: the first letter
> being uppercase).  I would be happy to work on that problem but
> it is a separate case and would need more discussion.

I agree with you, and it has been my plan as editor of ISO 30112 to add functionality along the lines that you suggest,
without having a firm propoal.

> Back about the main problem, I would like to post a patch which would
> be the best explanation what I mean in this particular case.

What would it say? 

keld

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-01 20:54       ` Keld Simonsen
@ 2019-10-02  7:33         ` Mikhail Gribanov
  2019-10-02  7:36           ` Mikhail Gribanov
  2019-10-02 22:14         ` Rafal Luzynski
  1 sibling, 1 reply; 14+ messages in thread
From: Mikhail Gribanov @ 2019-10-02  7:33 UTC (permalink / raw)
  To: Keld Simonsen; +Cc: Rafal Luzynski, libc-alpha, bug-glibc-locales

>Since the solution is controversial I would like to hear
>feedback from Mikhail and/or Keld (CC).  This patch does not
>determine whether lower or uppercase is correct, just says
>that ru_RU and ru_UA should be the same.

Just use the same representation for abday as in ru_RU as it is (the
first letter is large, the rest is small). It is correct.

I want ru_UA locale to inherit all translations from ru_RU (months,
days of the week, etc.) excluding government data (currency sign,
phone number, etc.). So that there is no such thing that the days of
the week in these locales are written differently (like now). They use
the same language, therefore the translation must be the same.

My native language is Russian and Ukrainian, I'm not a linguist, so
I’ll only express my point of view. Abbreviations for days of the week
(two letters) and month (three letters) in Russian and Ukrainian in
everyday life is always written with a first capital letter ("Пн",
"Вт"). But if this is due to some limitations - then use small letters
(but it is wrong). It’s easier for me, as an end user, to correct
format string in app than to correct a translation every time.

>You replied to me off-list.  I don't mind this but was this intentional?

My mistake, I use the web version of gmail, maybe pressed the wrong button.

вт, 1 окт. 2019 г. в 23:54, Keld Simonsen <keld@keldix.com>:
>
> On Tue, Oct 01, 2019 at 09:45:13PM +0200, Rafal Luzynski wrote:
> > 28.09.2019 19:37 Keld Simonsen <keld@keldix.com> wrote:
> > >
> > > I think we should use lowercase that is what we do with manuyothe european
> > > locales
> > > and it is the culturally correct form
> > >
> > > keld
> >
> > Thank you for your feedback, Keld.  Unfortunately, as a native speaker
> > of one of those languages I must say that this is more complex.
> > Indeed, names of the weeks and months are usually lowercase but
> > when they appear in the beginning of a sentence they are uppercase.
> > I am afraid that a name of a weekday more often appears in the beginning
> > of the sentence than in the middle.  To the extent that some locales
> > prefer the weekday and/or month names to start with the uppercase.
>
> I am also a native speaker of one of those languages, Danish, and moreover I was the initial contributer of about 60 of the locales incl ru_RU,
> that Urich Drepper picked up for glibc. I am also the editor of ISO 14652 and ISO 30112 that provided extensions
> beyond Posix i18n functionality, and I  provide guidance for their use. My advice is that this bad policy to have wrongly
> capitalized names here.
>
> > What we actually need is a new format modifier to control the first
> > letter being uppercase/lowercase.  Currently we have only "^" to convert
> > whole substring (e.g., a weekday name) to uppercase, and "#" to swap
> > the upper/lowercase.  It was discussed in the past, there is a bug
> > report and a proposal to ensure that "^#" works as a converter to
> > lowercase but this is not enough because I would like to see one more
> > format modifier to convert to titlecase (that is: the first letter
> > being uppercase).  I would be happy to work on that problem but
> > it is a separate case and would need more discussion.
>
> I agree with you, and it has been my plan as editor of ISO 30112 to add functionality along the lines that you suggest,
> without having a firm propoal.
>
> > Back about the main problem, I would like to post a patch which would
> > be the best explanation what I mean in this particular case.
>
> What would it say?
>
> keld

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-02  7:33         ` Mikhail Gribanov
@ 2019-10-02  7:36           ` Mikhail Gribanov
  0 siblings, 0 replies; 14+ messages in thread
From: Mikhail Gribanov @ 2019-10-02  7:36 UTC (permalink / raw)
  To: Keld Simonsen, Rafal Luzynski, libc-alpha, bug-glibc-locales

>Since the solution is controversial I would like to hear
>feedback from Mikhail and/or Keld (CC).  This patch does not
>determine whether lower or uppercase is correct, just says
>that ru_RU and ru_UA should be the same.

Just use the same representation for abday as in ru_RU as it is (the
first letter is large, the rest is small). It is correct.

I want ru_UA locale to inherit all translations from ru_RU (months,
days of the week, etc.) excluding government data (currency sign,
phone number, etc.). So that there is no such thing that the days of
the week in these locales are written differently (like now). They use
the same language, therefore the translation must be the same.

My native language is Russian and Ukrainian, I'm not a linguist, so
I’ll only express my point of view. Abbreviations for days of the week
(two letters) and month (three letters) in Russian and Ukrainian in
everyday life is always written with a first capital letter ("Пн",
"Вт"). But if this is due to some limitations - then use small letters
(but it is wrong). It’s easier for me, as an end user, to correct
format string in app than to correct a translation every time.

>You replied to me off-list.  I don't mind this but was this intentional?

My mistake, I use the web version of gmail, maybe pressed the wrong button.

ср, 2 окт. 2019 г. в 10:33, Mikhail Gribanov <mikhail.gribanov@gmail.com>:
>
> >Since the solution is controversial I would like to hear
> >feedback from Mikhail and/or Keld (CC).  This patch does not
> >determine whether lower or uppercase is correct, just says
> >that ru_RU and ru_UA should be the same.
>
> Just use the same representation for abday as in ru_RU as it is (the
> first letter is large, the rest is small). It is correct.
>
> I want ru_UA locale to inherit all translations from ru_RU (months,
> days of the week, etc.) excluding government data (currency sign,
> phone number, etc.). So that there is no such thing that the days of
> the week in these locales are written differently (like now). They use
> the same language, therefore the translation must be the same.
>
> My native language is Russian and Ukrainian, I'm not a linguist, so
> I’ll only express my point of view. Abbreviations for days of the week
> (two letters) and month (three letters) in Russian and Ukrainian in
> everyday life is always written with a first capital letter ("Пн",
> "Вт"). But if this is due to some limitations - then use small letters
> (but it is wrong). It’s easier for me, as an end user, to correct
> format string in app than to correct a translation every time.
>
> >You replied to me off-list.  I don't mind this but was this intentional?
>
> My mistake, I use the web version of gmail, maybe pressed the wrong button.
>
> вт, 1 окт. 2019 г. в 23:54, Keld Simonsen <keld@keldix.com>:
> >
> > On Tue, Oct 01, 2019 at 09:45:13PM +0200, Rafal Luzynski wrote:
> > > 28.09.2019 19:37 Keld Simonsen <keld@keldix.com> wrote:
> > > >
> > > > I think we should use lowercase that is what we do with manuyothe european
> > > > locales
> > > > and it is the culturally correct form
> > > >
> > > > keld
> > >
> > > Thank you for your feedback, Keld.  Unfortunately, as a native speaker
> > > of one of those languages I must say that this is more complex.
> > > Indeed, names of the weeks and months are usually lowercase but
> > > when they appear in the beginning of a sentence they are uppercase.
> > > I am afraid that a name of a weekday more often appears in the beginning
> > > of the sentence than in the middle.  To the extent that some locales
> > > prefer the weekday and/or month names to start with the uppercase.
> >
> > I am also a native speaker of one of those languages, Danish, and moreover I was the initial contributer of about 60 of the locales incl ru_RU,
> > that Urich Drepper picked up for glibc. I am also the editor of ISO 14652 and ISO 30112 that provided extensions
> > beyond Posix i18n functionality, and I  provide guidance for their use. My advice is that this bad policy to have wrongly
> > capitalized names here.
> >
> > > What we actually need is a new format modifier to control the first
> > > letter being uppercase/lowercase.  Currently we have only "^" to convert
> > > whole substring (e.g., a weekday name) to uppercase, and "#" to swap
> > > the upper/lowercase.  It was discussed in the past, there is a bug
> > > report and a proposal to ensure that "^#" works as a converter to
> > > lowercase but this is not enough because I would like to see one more
> > > format modifier to convert to titlecase (that is: the first letter
> > > being uppercase).  I would be happy to work on that problem but
> > > it is a separate case and would need more discussion.
> >
> > I agree with you, and it has been my plan as editor of ISO 30112 to add functionality along the lines that you suggest,
> > without having a firm propoal.
> >
> > > Back about the main problem, I would like to post a patch which would
> > > be the best explanation what I mean in this particular case.
> >
> > What would it say?
> >
> > keld

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-01 20:54       ` Keld Simonsen
  2019-10-02  7:33         ` Mikhail Gribanov
@ 2019-10-02 22:14         ` Rafal Luzynski
  2019-10-03  6:24           ` Mikhail Gribanov
                             ` (2 more replies)
  1 sibling, 3 replies; 14+ messages in thread
From: Rafal Luzynski @ 2019-10-02 22:14 UTC (permalink / raw)
  To: Keld Simonsen; +Cc: Mikhail Gribanov, libc-alpha, bug-glibc-locales

1.10.2019 22:54 Keld Simonsen <keld@keldix.com> wrote:
> [...]
> I am also a native speaker of one of those languages, Danish,

A test:

$ LANG=da_DK.utf8 date
tor  3 okt 00:00:05 CEST 2019
$ LANG=ru_RU.utf8 date
Чт окт  3 00:00:23 CEST 2019
$ LANG=pl_PL.utf8 date
czw, 3 paź 2019, 00:00:31 CEST

Do I understand correctly that "tor" is correct here?  If yes,
it is OK, I am not an expert at your language. :-)

I would say that Polish "czw" is not correct.  Indeed, normally
we write weekday names (and month names) lowercase but as this is
the beginning of a sentence the uppercase first letter would be
better.  For the same reason I think that Russian "Чт" is correct
even if CLDR says otherwise.

> and moreover I was the initial contributer of about 60 of the locales incl
> ru_RU,
> that Urich Drepper picked up for glibc. I am also the editor of ISO 14652
> and ISO 30112 that provided extensions
> beyond Posix i18n functionality, and I  provide guidance for their use.

I know and therefore I respect your opinion.

> My advice is that this bad policy to have wrongly
> capitalized names here. 

I think we don't have a consensus here and therefore I will not push
this commit unless someone convinces me to push it.

> [...]
> > Back about the main problem, I would like to post a patch which would
> > be the best explanation what I mean in this particular case.
> 
> What would it say? 

I think I posted it just while you were writing your message.
I will appreciate if you take a look.  It does not say clearly whether
upper/lowercase letters are correct, just that ru_UA should be a copy
of ru_RU.

Regards,

Rafal

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-02 22:14         ` Rafal Luzynski
@ 2019-10-03  6:24           ` Mikhail Gribanov
  2019-10-05  8:04           ` Maciej W. Rozycki
  2019-10-05  8:56           ` Keld Simonsen
  2 siblings, 0 replies; 14+ messages in thread
From: Mikhail Gribanov @ 2019-10-03  6:24 UTC (permalink / raw)
  To: Rafal Luzynski; +Cc: Keld Simonsen, libc-alpha, bug-glibc-locales

>A test:
>$ LANG=ru_RU.utf8 date
>Чт окт  3 00:00:23 CEST 2019
Actually, correct is "Чт, 3 окт 2019, 00:00:23 CEST". Android uses a
similar datetime format string.

In any case, I appeal to everyone. The discussion escalated from a
discussion of one small bug in discussions with which letters to use.
It is not so important for me. Big letters, small - no difference.
Ordinary users can fix it on their side. Date formatting in Linux is a
complete mess (at least in RU/UA locales), so manual intervention is
always required.

Please fix the translation bug first. I even offered you a solution:
https://sourceware.org/bugzilla/show_bug.cgi?id=25044
Copy abday values from ru_RU to ru_UA. That’s all I want. Uppercase or
lowercase discussion does not apply to this bug in any way.

Thank you for understanding.

чт, 3 окт. 2019 г. в 01:13, Rafal Luzynski <digitalfreak@lingonborough.com>:
>
> 1.10.2019 22:54 Keld Simonsen <keld@keldix.com> wrote:
> > [...]
> > I am also a native speaker of one of those languages, Danish,
>
> A test:
>
> $ LANG=da_DK.utf8 date
> tor  3 okt 00:00:05 CEST 2019
> $ LANG=ru_RU.utf8 date
> Чт окт  3 00:00:23 CEST 2019
> $ LANG=pl_PL.utf8 date
> czw, 3 paź 2019, 00:00:31 CEST
>
> Do I understand correctly that "tor" is correct here?  If yes,
> it is OK, I am not an expert at your language. :-)
>
> I would say that Polish "czw" is not correct.  Indeed, normally
> we write weekday names (and month names) lowercase but as this is
> the beginning of a sentence the uppercase first letter would be
> better.  For the same reason I think that Russian "Чт" is correct
> even if CLDR says otherwise.
>
> > and moreover I was the initial contributer of about 60 of the locales incl
> > ru_RU,
> > that Urich Drepper picked up for glibc. I am also the editor of ISO 14652
> > and ISO 30112 that provided extensions
> > beyond Posix i18n functionality, and I  provide guidance for their use.
>
> I know and therefore I respect your opinion.
>
> > My advice is that this bad policy to have wrongly
> > capitalized names here.
>
> I think we don't have a consensus here and therefore I will not push
> this commit unless someone convinces me to push it.
>
> > [...]
> > > Back about the main problem, I would like to post a patch which would
> > > be the best explanation what I mean in this particular case.
> >
> > What would it say?
>
> I think I posted it just while you were writing your message.
> I will appreciate if you take a look.  It does not say clearly whether
> upper/lowercase letters are correct, just that ru_UA should be a copy
> of ru_RU.
>
> Regards,
>
> Rafal

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-02 22:14         ` Rafal Luzynski
  2019-10-03  6:24           ` Mikhail Gribanov
@ 2019-10-05  8:04           ` Maciej W. Rozycki
  2019-10-05 19:58             ` Rafal Luzynski
  2019-10-05  8:56           ` Keld Simonsen
  2 siblings, 1 reply; 14+ messages in thread
From: Maciej W. Rozycki @ 2019-10-05  8:04 UTC (permalink / raw)
  To: Rafal Luzynski
  Cc: Keld Simonsen, Mikhail Gribanov, libc-alpha, bug-glibc-locales

On Thu, 3 Oct 2019, Rafal Luzynski wrote:

> $ LANG=pl_PL.utf8 date
> czw, 3 paź 2019, 00:00:31 CEST
> 
> Do I understand correctly that "tor" is correct here?  If yes,
> it is OK, I am not an expert at your language. :-)
> 
> I would say that Polish "czw" is not correct.  Indeed, normally
> we write weekday names (and month names) lowercase but as this is
> the beginning of a sentence the uppercase first letter would be
> better.  For the same reason I think that Russian "Чт" is correct
> even if CLDR says otherwise.

 Hmm, can a timestamp be considered a sentence or a nominal sentence?

 Moreover, is the use of three letters as the abbreviation correct in the 
first place regardless of its capitalisation?  Common usage as in printed 
calendars for the days of the week in Polish has been (starting from Mon): 
Pn, Wt, Śr, Cz, Pt, So, N (capitalised) and what we have in locale appears 
pretty arbitrary to me, e.g. "nie" does not appear defined by any language 
standard, with "niedz." or "nd." (with a full stop necessarily included) 
being the acceptable abbreviations in written text other than calendars.

 Maybe I'm missing something, but I have always found the representation 
of time stamps in our Polish locale disturbingly odd (and, alas, I do not 
use it).

 FWIW,

  Maciej

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-02 22:14         ` Rafal Luzynski
  2019-10-03  6:24           ` Mikhail Gribanov
  2019-10-05  8:04           ` Maciej W. Rozycki
@ 2019-10-05  8:56           ` Keld Simonsen
  2019-10-05 20:16             ` Rafal Luzynski
  2 siblings, 1 reply; 14+ messages in thread
From: Keld Simonsen @ 2019-10-05  8:56 UTC (permalink / raw)
  To: Rafal Luzynski; +Cc: Mikhail Gribanov, libc-alpha, bug-glibc-locales

On Thu, Oct 03, 2019 at 12:12:49AM +0200, Rafal Luzynski wrote:
> 1.10.2019 22:54 Keld Simonsen <keld@keldix.com> wrote:
> > [...]
> > I am also a native speaker of one of those languages, Danish,
> 
> A test:
> 
> $ LANG=da_DK.utf8 date
> tor  3 okt 00:00:05 CEST 2019
> $ LANG=ru_RU.utf8 date
> ???? ??????  3 00:00:23 CEST 2019
> $ LANG=pl_PL.utf8 date
> czw, 3 pa?? 2019, 00:00:31 CEST
> 
> Do I understand correctly that "tor" is correct here?  If yes,
> it is OK, I am not an expert at your language. :-)

For Danish, this is fine, for listings and logs etc.
The beginning of a sentence is normally capitalized in Danish, but this is not a sentence.
And having the day an month names capitalized leads confusion, you see a lot of occurrances
in just plain text, where these names  are capitalized - and people believe ths is correct, also because 
this is what is done in English.

Also I am associated with the official Danish standarsization organisation, where I need to state Danish info in accordance
with official conventions. and capitalizing the first letter because it would normally occur in the beginning
of a sentence is IMHO a hack.

using 3-letter abbreviatons is posix style. In danish I think 2-letter are more cmmon, but the 3-letters are wel understood, as you
in 5 cases just need to add 'dag' to get the full name of the day: man, ons, fre, lør, søn, for  month names 2 letters are not enough but
3 letters are fine in Danish.

I believe the abday and abmon specs are for use in posix utilities, so they should  be posix-style, eg. all same length so
file lisings with ls, and log entries format nicely.
For other uses of more cultural oriented we should add new keywords, like maybe abday_cult
> 
> I would say that Polish "czw" is not correct.  Indeed, normally
> we write weekday names (and month names) lowercase but as this is
> the beginning of a sentence the uppercase first letter would be
> better.  For the same reason I think that Russian "????" is correct
> even if CLDR says otherwise.


I would rather solve this with a formatter spec to capitalice the first letter
for example month names would nomally not be the first word in a sentence, an often the day-of-month would precede it.
capitalizing day names and not month names - against the rules of the language, is IMHO a mess.

> > and moreover I was the initial contributer of about 60 of the locales incl
> > ru_RU,
> > that Urich Drepper picked up for glibc. I am also the editor of ISO 14652
> > and ISO 30112 that provided extensions
> > beyond Posix i18n functionality, and I  provide guidance for their use.
> 
> I know and therefore I respect your opinion.
> 
> > My advice is that this bad policy to have wrongly
> > capitalized names here. 
> 
> I think we don't have a consensus here and therefore I will not push
> this commit unless someone convinces me to push it.
> 
> > [...]
> > > Back about the main problem, I would like to post a patch which would
> > > be the best explanation what I mean in this particular case.
> > 
> > What would it say? 
> 
> I think I posted it just while you were writing your message.
> I will appreciate if you take a look.  It does not say clearly whether
> upper/lowercase letters are correct, just that ru_UA should be a copy
> of ru_RU.

 I saw your advice and I agree. Just do the copy.

Keld

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-05  8:04           ` Maciej W. Rozycki
@ 2019-10-05 19:58             ` Rafal Luzynski
  0 siblings, 0 replies; 14+ messages in thread
From: Rafal Luzynski @ 2019-10-05 19:58 UTC (permalink / raw)
  To: Maciej W. Rozycki
  Cc: Keld Simonsen, Mikhail Gribanov, libc-alpha, bug-glibc-locales

5.10.2019 10:04 "Maciej W. Rozycki" <macro@linux-mips.org> wrote:
> 
> On Thu, 3 Oct 2019, Rafal Luzynski wrote:
> 
> > $ LANG=pl_PL.utf8 date
> > czw, 3 paź 2019, 00:00:31 CEST
> > 
> > Do I understand correctly that "tor" is correct here?  If yes,
> > it is OK, I am not an expert at your language. :-)
> > 
> > I would say that Polish "czw" is not correct.  Indeed, normally
> > we write weekday names (and month names) lowercase but as this is
> > the beginning of a sentence the uppercase first letter would be
> > better.  For the same reason I think that Russian "Чт" is correct
> > even if CLDR says otherwise.
> 
>  Hmm, can a timestamp be considered a sentence or a nominal sentence?

In Polish language, no matter if it is a nominal or a verbal sentence,
it should be capitalized.  I assume that for Russian as well.  But it is
a minor error if we do not capitalize.

>  Moreover, is the use of three letters as the abbreviation correct in the 
> first place regardless of its capitalisation?  Common usage as in printed 
> calendars for the days of the week in Polish has been (starting from Mon):
> 
> Pn, Wt, Śr, Cz, Pt, So, N (capitalised) and what we have in locale appears
> 
> pretty arbitrary to me, e.g. "nie" does not appear defined by any language
> 
> standard, with "niedz." or "nd." (with a full stop necessarily included) 
> being the acceptable abbreviations in written text other than calendars.

AFAIK there is no official specification of the abbreviated forms in the
Polish language so any abbreviation is correct.  Also, in Glibc we
strongly rely on CLDR. [1] It provides more forms that we need, that is
both exactly three letter abbreviations and some longer abbreviations like
"niedz."  On the other hand, I don't think it is necessary to provide longer
forms, there are some tools (e.g., cal) which truncate abbreviated weekday
names.

>  Maybe I'm missing something, but I have always found the representation 
> of time stamps in our Polish locale disturbingly odd (and, alas, I do not 
> use it).

I'll be happy to fix it but, again, so far I am not aware of any issue
where CLDR provides other data than we have in Glibc (in Polish locale).

Regards,

Rafal


[1] https://st.unicode.org/cldr-apps/v#/pl/Gregorian/

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: Incorrect days of the week in ru_UA locale
  2019-10-05  8:56           ` Keld Simonsen
@ 2019-10-05 20:16             ` Rafal Luzynski
  0 siblings, 0 replies; 14+ messages in thread
From: Rafal Luzynski @ 2019-10-05 20:16 UTC (permalink / raw)
  To: Keld Simonsen; +Cc: Mikhail Gribanov, libc-alpha, bug-glibc-locales

5.10.2019 10:56 Keld Simonsen <keld@keldix.com> wrote:
> On Thu, Oct 03, 2019 at 12:12:49AM +0200, Rafal Luzynski wrote:
> > [...]
> > Do I understand correctly that "tor" is correct here?  If yes,
> > it is OK, I am not an expert at your language. :-)
> 
> For Danish, this is fine, for listings and logs etc.
> The beginning of a sentence is normally capitalized in Danish, but this is
> not a sentence.

OK, I trust you.

> And having the day an month names capitalized leads confusion, you see a
> lot of occurrances
> in just plain text, where these names  are capitalized

No, I did not mean to capitalize both day and month names.  I only
meant to capitalize whatever is first, weekday names in this case.
And, again, as you can see, only in some languages.

> - and people believe ths is correct, also because 
> this is what is done in English.

I always care about the ability to apply the local language rules.
Applying English rules to other languages is (often) wrong.

> Also I am associated with the official Danish standarsization
> organisation, where I need to state Danish info in accordance
> with official conventions. and capitalizing the first letter because it
> would normally occur in the beginning
> of a sentence is IMHO a hack.
> 
> using 3-letter abbreviatons is posix style.

I think we had a similar discussion in the past and it has been said
that POSIX does not say anything about other languages than English.
Therefore other languages may use different rules if it is correct
for them.

> In danish I think 2-letter are more cmmon, but the 3-letters are wel
> understood, as you
> in 5 cases just need to add 'dag' to get the full name of the day: man,
> ons, fre, lør, søn, for  month names 2 letters are not enough but
> 3 letters are fine in Danish.
> 
> I believe the abday and abmon specs are for use in posix utilities, so
> they should  be posix-style, eg. all same length so
> file lisings with ls, and log entries format nicely.
> For other uses of more cultural oriented we should add new keywords, like
> maybe abday_cult

In many locales we use other abbreviations than 3 letters:
shorter, longer, varying.  Command line utilities somehow deal with this,
some just truncate the output, some ensure enough space for all possible
lengths etc.

> > I would say that Polish "czw" is not correct.  Indeed, normally
> > we write weekday names (and month names) lowercase but as this is
> > the beginning of a sentence the uppercase first letter would be
> > better.  For the same reason I think that Russian "????" is correct
> > even if CLDR says otherwise.
> 
> 
> I would rather solve this with a formatter spec to capitalice the first
> letter
> for example month names would nomally not be the first word in a sentence,
> an often the day-of-month would precede it.
> [...]

Sure, that's what I mean.  We can work on this in the future.

> > [...]
> > I think I posted it just while you were writing your message.
> > I will appreciate if you take a look.  It does not say clearly whether
> > upper/lowercase letters are correct, just that ru_UA should be a copy
> > of ru_RU.
> 
>  I saw your advice and I agree. Just do the copy.

That's ambiguous: do you mean I should:

1. Use “copy "ru_RU"” in ru_UA as I suggested in my patch?
or
2. Copy & paste (in an editor) abday from ru_RU to ru_UA?

Both solve the main issue except that the second one leaves a difference
in ab_alt_mon between ru_RU and ru_UA, the difference in capitalization
only.  Which became the issue of this discussion.

Regards,

Rafal

^ permalink raw reply	[flat|nested] 14+ messages in thread

end of thread, other threads:[~2019-10-05 20:16 UTC | newest]

Thread overview: 14+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-09-27 15:25 Incorrect days of the week in ru_UA locale Mikhail Gribanov
2019-09-27 15:44 ` Florian Weimer
2019-09-27 22:49 ` Rafal Luzynski
2019-09-28 17:25   ` Keld Simonsen
2019-10-01 19:45     ` Rafal Luzynski
2019-10-01 20:54       ` Keld Simonsen
2019-10-02  7:33         ` Mikhail Gribanov
2019-10-02  7:36           ` Mikhail Gribanov
2019-10-02 22:14         ` Rafal Luzynski
2019-10-03  6:24           ` Mikhail Gribanov
2019-10-05  8:04           ` Maciej W. Rozycki
2019-10-05 19:58             ` Rafal Luzynski
2019-10-05  8:56           ` Keld Simonsen
2019-10-05 20:16             ` Rafal Luzynski

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).