Here is my "solution" to the encoding problem that a lot of people have been complaining about. The problem is that d library functions cannot read or write anything but unicode from stdin/stdout. So for fun I made:

http://folk.uio.no/mortennk/d/locale_v0.1.zip

Quick usage example:

# import locale.translate;
#
# void main()
# {
#    // Read a line from stdin, it is converted to utf-8.
#    char[] a = tin.readLine();
#
#    // a is now unicode, you can do what you want with it here
#    // ...
#
#    // Write the string out again, converting it automatically back to the
#    // local charset.
#    tout.writefln("Your string is ", a);
# }

(The file also contains some general character conversion routines.)

Ok, it's not meant as a perfect solution, or even a good one. Just to demonstrate _one_ way to solve the problem.

Oh, and it uses iconv and is currently linux-only. Sorry :) Any feedback is appreciated.

Nick

Nick wrote:

> Here is my "solution" to the encoding problem that a lot of people have been
> complaining about. The problem is that d library functions cannot read or write
> anything but unicode from stdin/stdout. So for fun I made:
[snip]
> Oh, and it uses iconv and is currently linux-only. Sorry :) Any feedback is
> appreciated.

Here is a more portable way of using iconv:
http://www.algonet.se/~afb/d/libiconv.d

Still, the iconv library is kinda huge ?
> 968K    libiconv.2.dylib
> 872K    iconv.dll
Even if it is built-in on most posix systems.

Good for the general case, but sorta overkill
for Latin-1 and other simple 8-bit encodings ?


And then I haven't even begun on locale... ;-)
(the code you posted uses GNU glibc version)

You see, the BSD locale.h has these instead:
> #define LC_ALL          0
> #define LC_COLLATE      1
> #define LC_CTYPE        2
> #define LC_MONETARY     3
> #define LC_NUMERIC      4
> #define LC_TIME         5
> #define LC_MESSAGES     6

And that's even more fun to port than the subtle
differences between GNU libiconv and BSD iconv...

--anders

"Anders F Björklund" <afb@algonet.se> wrote in message news:d04ce0$2nis$1@digitaldaemon.com...
> Nick wrote:
>
>> Here is my "solution" to the encoding problem that a lot of people have
>> been
>> complaining about. The problem is that d library functions cannot read or
>> write
>> anything but unicode from stdin/stdout. So for fun I made:
> [snip]
>> Oh, and it uses iconv and is currently linux-only. Sorry :) Any feedback
>> is
>> appreciated.
>
> Here is a more portable way of using iconv: http://www.algonet.se/~afb/d/libiconv.d

Ah - cool! I had forgotten you had done that. Can we put the link in a FAQ somewhere? Together with a link to Mango's ICU port. Plus examples on how to use them.

By the way, how many FAQs do we have anyway? I see http://www.digitalmars.com/d/faq.html http://www.prowiki.org/wiki4d/wiki.cgi?FaqRoadmap

In article <d04lhg$oe$1@digitaldaemon.com>, Ben Hinkle says...
>
>By the way, how many FAQs do we have anyway? I see http://www.digitalmars.com/d/faq.html http://www.prowiki.org/wiki4d/wiki.cgi?FaqRoadmap
>

Ben: Well I'd say we have at least 4 FAQs, if you add the following two:

DedicateD (http://int19h.tamb.ru/faq.html)
MKoD (my D site)(http://spottedtiger.tripod.com/D_Language/D_FAQ_XP.html)

David L.

-------------------------------------------------------------------
"Dare to reach for the Stars...Dare to Dream, Build, and Achieve!"

Ben Hinkle wrote:

>>Here is a more portable way of using iconv:
>>http://www.algonet.se/~afb/d/libiconv.d
> 
> Ah - cool! I had forgotten you had done that. Can we put the link in a FAQ somewhere? Together with a link to Mango's ICU port. Plus examples on how to use them.

I throw a bunch of hacks on my personal site,
that I don't really know where else to put...
(or didn't have the time to clean up for release)

I'll be happy to contribute some elsewhere ?

--anders