R devel - Jun 2019 - Setting LC_CTYPE=en_US.UTF-8 failed

On 06/06/2019 6:22 a.m., Tomas Kalibera wrote:
> On 6/5/19 3:49 AM, Steven Penny wrote:
>> Using this in my "~/.profile":
>>
>>  ?? export LC_ALL=en_US.UTF-8
>>
>> Yields this:
>>
>>  ?? $ Rscript -e 'print(9)'
>>  ?? During startup - Warning message:
>>  ?? Setting LC_CTYPE=en_US.UTF-8 failed
>>  ?? [1] 9
>>
>> This is confusing as the exact same environment works fine with other
>> languages:
>>
>>  ?? $ python3 -c 'print(9)'
>>  ?? 9
>>
>>  ?? $ ruby -e 'puts 9'
>>  ?? 9
> 
> The locale is probably not available on your system, please
> install/generate it.
I think Steven is running the Windows build, not a POSIX build, so 
that's not a legal value.  He could try the Cygwin build, but I think it 
has other problems, and isn't supported by R Core.  As far as I know 
there is no Windows locale that fully supports UTF-8.

Duncan Murdoch

On 06/06/2019 7:28 a.m., Duncan Murdoch wrote:
> On 06/06/2019 6:22 a.m., Tomas Kalibera wrote:
>> On 6/5/19 3:49 AM, Steven Penny wrote:
>>> Using this in my "~/.profile":
>>>
>>>   ?? export LC_ALL=en_US.UTF-8
>>>
>>> Yields this:
>>>
>>>   ?? $ Rscript -e 'print(9)'
>>>   ?? During startup - Warning message:
>>>   ?? Setting LC_CTYPE=en_US.UTF-8 failed
>>>   ?? [1] 9
>>>
>>> This is confusing as the exact same environment works fine with
other
>>> languages:
>>>
>>>   ?? $ python3 -c 'print(9)'
>>>   ?? 9
>>>
>>>   ?? $ ruby -e 'puts 9'
>>>   ?? 9
>>
>> The locale is probably not available on your system, please
>> install/generate it.
> 
> I think Steven is running the Windows build, not a POSIX build, so
> that's not a legal value.  He could try the Cygwin build, but I think
it
> has other problems, and isn't supported by R Core.  As far as I know
> there is no Windows locale that fully supports UTF-8.
> 
The Windows name for a UTF-8 locale is supposed to be en_US.65001 (see
<https://docs.microsoft.com/en-us/cpp/c-runtime-library/locale-names-languages-and-country-region-strings?view=vs-2019>
if it hasn't disappeared by the time you read this).  However, when I 
tried that a few years ago, it didn't work.  Maybe it does now.

Duncan Murdoch

On 6/6/19 3:24 PM, Duncan Murdoch wrote:
> On 06/06/2019 7:28 a.m., Duncan Murdoch wrote:
>> On 06/06/2019 6:22 a.m., Tomas Kalibera wrote:
>>> On 6/5/19 3:49 AM, Steven Penny wrote:
>>>> Using this in my "~/.profile":
>>>>
>>>> ? ?? export LC_ALL=en_US.UTF-8
>>>>
>>>> Yields this:
>>>>
>>>> ? ?? $ Rscript -e 'print(9)'
>>>> ? ?? During startup - Warning message:
>>>> ? ?? Setting LC_CTYPE=en_US.UTF-8 failed
>>>> ? ?? [1] 9
>>>>
>>>> This is confusing as the exact same environment works fine with
other
>>>> languages:
>>>>
>>>> ? ?? $ python3 -c 'print(9)'
>>>> ? ?? 9
>>>>
>>>> ? ?? $ ruby -e 'puts 9'
>>>> ? ?? 9
>>>
>>> The locale is probably not available on your system, please
>>> install/generate it.
>>
>> I think Steven is running the Windows build, not a POSIX build, so
>> that's not a legal value.? He could try the Cygwin build, but I
think it
>> has other problems, and isn't supported by R Core.? As far as I
know
>> there is no Windows locale that fully supports UTF-8.
>>
>
> The Windows name for a UTF-8 locale is supposed to be en_US.65001 (see
>
<https://docs.microsoft.com/en-us/cpp/c-runtime-library/locale-names-languages-and-country-region-strings?view=vs-2019>
> if it hasn't disappeared by the time you read this).? However, when I 
> tried that a few years ago, it didn't work.? Maybe it does now.
Yes, it is 65001 but it still cannot be used as setlocale() Windows 
locale (in LC_CTYPE), and hence R on Windows cannot use UTF-8 as native 
encoding.

Tomas
>
> Duncan Murdoch