Mkdir and utf8

perlmonkdr has asked for the wisdom of the Perl Monks concerning the following question:

Hello Monks,

I'm again, :-), i can believe that one step is one problem to me, well i was tried to create a directory with utf8 name encoding but can do this, i'm read some question similar in this thread 604999, can't found any solution.

Check this code, save in utf-8:

#!/usr/bin/perl

use utf8;

mkdir "Ïntêràtíonaliçzation";

system "mkdir Ïntêràtíonal";

open F,'>:utf8','./test.txt';
print F "Ïntêràtíonaliçzation";
close F;
[download]

If you see the test.txt file the word is ok encoded, but both makes dir not at least in win xp.

How can i tell to perl that create the directory in utf8

Many thank

Comment on Mkdir and utf8 Download Code

Replies are listed 'Best First'.
Re: Mkdir and utf8 by moritz (Cardinal) on Nov 07, 2007 at 18:36 UTC
Try this: `use utf8; use Encode qw/encode/; mkdir encode('utf-8', "Ïntêràtíonaliçzation");` [download] (works for me on linux)	[reply] [d/l]
Re^2: Mkdir and utf8 by perlmonkdr (Beadle) on Nov 08, 2007 at 00:16 UTC
Thk U moritz, is the second time that you help me, in this case this example not works on windows, if you need, check the response below, that works ok. Thk U again	[reply]
Re: Mkdir and utf8 by ikegami (Patriarch) on Nov 07, 2007 at 21:05 UTC
Windows has two set of system calls: "A"NSI and "W"ide. Perl's file functions use the "A" version of the system call, so the UTF-8 encoding of the string you pass to them is interpreted as being encoded according to the system's code page (e.g. American systems typically use `cp1252`, Microsoft's version of `iso-latin-1`). One way to achieve your goal is to encode the file name according to the system's code page. `#!/usr/bin/perl use utf8; use Encode qw( encode ); use Win32::Codepage; mkdir encode(Win32::Codepage::get_encoding(), 'Ïntêràtíonaliçzation');` [download] That method assumes the characters you are using are present in your system's code page. If they are not, you will need to use the "W" version of the function. The "W" function accepts UCS-2le strings which can encode common UNICODE characters. Unfortunately, I don't think there currently exists a module that gives access to those functions, but you could access them via Win32::API.	[reply] [d/l] [select]
Re^2: Mkdir and utf8 by perlmonkdr (Beadle) on Nov 08, 2007 at 00:11 UTC
You right, the example that you give to us works exactly as you say on windows,i'm really not know that windows has two systems calls, i appreciate very much, the problem was finaly solved, i will also investigate the wide call. Thk U very much ikegami	[reply]
Re^3: Mkdir and utf8 by ikegami (Patriarch) on Nov 08, 2007 at 01:29 UTC
For example, the system call in question is `CreateDirectory`. `BOOL WINAPI CreateDirectory( LPCTSTR lpPathName, LPSECURITY_ATTRIBUTES lpSecurityAttributes );` [download] If you read down that page, you'll see that it's "implemented as CreateDirectoryW (Unicode) and CreateDirectoryA (ANSI)". `CreateDirectory` is really just an alias for one of these two actual KERNEL32 functions. If you look in `winbase.h`, you'll find `WINBASEAPI BOOL WINAPI CreateDirectoryA( LPCSTR lpPathName, LPSECURITY_ATTRIBUTES lpSecurityAttributes ); WINBASEAPI BOOL WINAPI CreateDirectoryW( LPCWSTR lpPathName, LPSECURITY_ATTRIBUTES lpSecurityAttributes ); #ifdef UNICODE #define CreateDirectory CreateDirectoryW #else #define CreateDirectory CreateDirectoryA #endif` [download]	[reply] [d/l] [select]
Re^3: Mkdir and utf8 by Anonymous Monk on Nov 08, 2007 at 07:48 UTC
Win32API::File - Low-level access to Win32 system API calls for files/dirs.	[reply]
Re: Mkdir and utf8 by apl (Monsignor) on Nov 07, 2007 at 18:22 UTC
Please, please, please test your returns! `mkdir "Ïntêràtíonaliçzation" \|\| die "mkdir failed $1";` Ditto your open...	[reply] [d/l]
Re^2: Mkdir and utf8 by ikegami (Patriarch) on Nov 07, 2007 at 19:09 UTC
You might find your solution (`mkdir("Ïntêràtíonaliçzation" \|\| die "mkdir failed $1")`) less than useful. Fix: `mkdir "Ïntêràtíonaliçzation" or die "mkdir failed $!";` [download]	[reply] [d/l] [select]
Re^3: Mkdir and utf8 by perlmonkdr (Beadle) on Nov 08, 2007 at 00:28 UTC
You are in all details!!, really thank and your votes were given. :-)	[reply]
Re^2: Mkdir and utf8 by perlmonkdr (Beadle) on Nov 08, 2007 at 00:23 UTC
Thk apl, perhaps i'm not be clear, usually it happens :-), the problem exists on character set, not in a fail on create dirs, anyway thank for your concern	[reply]
Re^3: Mkdir and utf8 by apl (Monsignor) on Nov 08, 2007 at 10:45 UTC
No, you were clear. I was anal-retentive, but as ikegami pointed out, incorrectly so.	[reply]