c++0x and locale_t

c++0x and locale_t

[Ken Brown]

I'm having an issue building icu, which boils down to the following test 
case:

$ cat foo.cc
#include <locale.h>
locale_t foo;

$ g++ -c --std=c++0x foo.cc
foo.cc:2:1: error: ‘locale_t’ does not name a type
  locale_t foo;
  ^

If I remove '--std=c++0x', the error goes away.  I know nothing about 
C++ standards, so I don't know if this is expected behavior or if it 
indicates a bug in Cygwin's headers.

Ken

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: c++0x and locale_t

[Brian Inglis]

On 2016-10-01 07:30, Ken Brown wrote:
> I'm having an issue building icu, which boils down to the following
> test case:
> $ cat foo.cc
> #include <locale.h>
> locale_t foo;
> $ g++ -c --std=c++0x foo.cc
> foo.cc:2:1: error: ‘locale_t’ does not name a type
>  locale_t foo;
>  ^
> If I remove '--std=c++0x', the error goes away.  I know nothing about
> C++ standards, so I don't know if this is expected behavior or if it
> indicates a bug in Cygwin's headers.

For C POSIX locale_t support, you have to do one or both of:

#define _XOPEN_SOURCE   700
#define _POSIX_C_SOURCE 200809L

to support multiple dynamic C locales and related functions.
This may be done automatically if you use the default -std=gnu++03, which may
have been the intent in ICU and original interpretation by g++.
g++ now interprets (and deprecates) c++0x to mean c++11.
You could try changing it to explicitly c++03 and see if it works, without the
GNU extensions.
Otherwise you should change it to explicitly gnu++03, as c++0x is deprecated,
and may be dropped; g++ also deprecates c++1y aka c++14 and c++1z which may be
c++17, and their gnu++ counterparts.

I don't understand why ICU C++ would use C locales, when C is now trying to add
a subset of features C++ has supported better, more flexibly in <locale> for
over a decade; see:

https://sourceforge.net/p/msys2/discussion/general/thread/23e1b5ce/

for a similar problem to yours, and the solution in standard C++; and:

http://stdcxx.apache.org/doc/stdlibug/24-3.html

for an explanation of the differences between C++ and C locales.

OTOH ICU comes from IBM, and may be more interested in consistency across
languages: how else can you explain C++ methods called createInstance?

But you may just be the packager, porter, and builder, so may be unable to fix
the implementation.

-- 
Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: c++0x and locale_t

[Ken Brown]

On 10/1/2016 10:58 AM, Brian Inglis wrote:
> On 2016-10-01 07:30, Ken Brown wrote:
>> I'm having an issue building icu, which boils down to the following
>> test case:
>> $ cat foo.cc
>> #include <locale.h>
>> locale_t foo;
>> $ g++ -c --std=c++0x foo.cc
>> foo.cc:2:1: error: ‘locale_t’ does not name a type
>>  locale_t foo;
>>  ^
>> If I remove '--std=c++0x', the error goes away.  I know nothing about
>> C++ standards, so I don't know if this is expected behavior or if it
>> indicates a bug in Cygwin's headers.
>
> For C POSIX locale_t support, you have to do one or both of:
>
> #define _XOPEN_SOURCE   700
> #define _POSIX_C_SOURCE 200809L
>
> to support multiple dynamic C locales and related functions.
> This may be done automatically if you use the default -std=gnu++03,
> which may
> have been the intent in ICU and original interpretation by g++.
> g++ now interprets (and deprecates) c++0x to mean c++11.
> You could try changing it to explicitly c++03 and see if it works,
> without the
> GNU extensions.
> Otherwise you should change it to explicitly gnu++03, as c++0x is
> deprecated,
> and may be dropped; g++ also deprecates c++1y aka c++14 and c++1z which
> may be
> c++17, and their gnu++ counterparts.
>
> I don't understand why ICU C++ would use C locales, when C is now trying
> to add
> a subset of features C++ has supported better, more flexibly in <locale>
> for
> over a decade; see:
>
> https://sourceforge.net/p/msys2/discussion/general/thread/23e1b5ce/
>
> for a similar problem to yours, and the solution in standard C++; and:
>
> http://stdcxx.apache.org/doc/stdlibug/24-3.html
>
> for an explanation of the differences between C++ and C locales.
>
> OTOH ICU comes from IBM, and may be more interested in consistency across
> languages: how else can you explain C++ methods called createInstance?
>
> But you may just be the packager, porter, and builder, so may be unable
> to fix
> the implementation.

Thanks for the information.  I've submitted a patch upstream.

Ken


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple