Re: [Boost-users] [filesystem] i18n merge complete

From: Beman Dawes (bdawes_at_[hidden])
Date: 2005-12-17 10:04:41

"Edward Diener" <eddielee_at_[hidden]> wrote in message 
news:do04tt$bvd$1_at_sea.gmane.org...
> Beman Dawes wrote:
>> The Boost.Filesystem i18n branch has been merged into the main trunk
>> HEAD in CVS.
>>
>> This is a major upgrade, several years in the making. It supports wide
>> character filenames, and has a pile of other improvements. It implements
>> the TR2 library proposal before the C++ Standards Committee.
>>
>> Most, but not all, current code using the library should continue to
>> work OK.
>>
>> Because of all the code changes, there may be teething troubles. I'll
>> try to watch the mailing lists and regression tests closely for the next
>> couple of weeks, and respond to problems ASAP.
>
> As someone who has argued vociferously on comp.std.c++ for wide
> character filenames, on the grounds that implementations which can map
> wide characters to Unicode code points can now support Unicode file
> names for appropriate OSs, I would like to thank you for adding the i18n
> support to the filesystem library.

Thanks! Postings on comp.std.c++ and Boost mailing lists were part of the 
motivation.

> I realize this does not directly add Unicode support or Unicode file
> names to C++, which is a separate issue, but on systems, such as
> Windows, where an end-user can switch to a code page for a language
> which needs Unicode support to express all characters, using your
> library should now enable them to create filenames using Unicode which
> are native to their own language.

Quite a lot of POSIX systems also support Unicode filenames. I've got 
filenames with little smiley faces and that sort of thing on my mini-mac.

> I do hope some Chinese, Japanese, or
> other Unicode language programmers will test this out and see if it is so.

I do to. While the wide-character filename test cases work OK, I'd love to 
get some usage reports from programmers using non-Latin alphabets.

--Beman