Re: [Cppcms-users] booster::date_time parsing

[Pavel K. <un...@fu...>]

On Wednesday 16 March 2011, Artyom wrote:
> http://cppcms.sourceforge.net/boost_locale/html/using_localization_backends
> .html
> 
> >std::string tz = "Europe/Prague";
> 
> Note you can use this:
> 
> http://cppcms.sourceforge.net/boost_locale/html/namespaceboost_1_1locale_1_
> 1time__zone.html
> To set global time zone.
> Also you can set timezone independently on stream using as::time_zone
> manipulator.

yes, they are very usefull. generally, i'm trying to write an example for 
generating current server timestamp, generating time from user input and 
storing it as a number (to DB) & restore it to timezone specific date & show 
it to user. maybe it will be usefull to have this in documentation (with best 
practise how to set it up with cppcms response()). i missed "as::time_zone" 
and ">>" parsing possibility for the first time while going through 
documentation.

 > It seems that you need to update the ICU version.
> I've tested this code (actually boost not booster, but it is the same):
> #include <boost/locale.hpp>
> #include <iostream>
> int main()
> {
>         std::locale::global(boost::locale::generator().generate(""));
>         std::stringstream s;
>         time_t d;
>         s.str("1955-07-21 07:21:07");
>         s >> boost::locale::as::ftime("%Y-%m-%d %H:%M:%S") >> d;
>         std::cout.imbue(std::locale());
>         std::cout << s.str() << std::endl;
>         std::cout << boost::locale::as::ftime("%Y-%m-%d %H:%M:%S") << d <<
> std::endl;
> }
>
> It hadn't work for icu 3.8 (gave a year 1970) but works without problems
> for icu-4.4 and 4.6,
> So it seems like ICU issue. (They have bugs as well :-( )
> 
> It is fairly simple to compile latest ICU version.

i work with icu-4.6 (gentoo package), and now tried it with 4.4.2, but without 
success. anyway, your example works as expected, so i figure out that 
 
s >> boost::locale::as::ftime("%Y-%m-%d %H:%M:%S") >>  _time_t_
versus
s >> boost::locale::as::ftime("%Y-%m-%d %H:%M:%S") >>  _date_time_

is the problem. thank you :).