localize problem

[vanchutr]

I'd ran Puppy 2.12
I'd used the Choselocale in Control Panel and set "locale to" en_US.UF-8 (I had the directory en_US.UTF-8 in two places: /usr/share/lib and /usr/share/locale.
When I use the command:
# locale
I met this message:

sh-3.00# locale
locale: Cannot set LC_CTYPE to default locale: No such file or directory
locale: Cannot set LC_ALL to default locale: No such file or directory
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
LC_MESSAGES="en_US.UTF-8"
LC_PAPER="en_US.UTF-8"
LC_NAME="en_US.UTF-8"
LC_ADDRESS="en_US.UTF-8"
LC_TELEPHONE="en_US.UTF-8"
LC_MEASUREMENT="en_US.UTF-8"
LC_IDENTIFICATION="en_US.UTF-8"
LC_ALL=
sh-3.00#

What errors did I met?
What can I do now? Please explain it to me and give me some instructions.
Thank a lot

[Ian]

I downloaded the locale files and could not find "en_US.UF-8", where did you get this from.

[vanchutr]

You can download here

[Ian]

Did you follow the instructions in the Chooselocale program as it explains where the files go and how to reverse the process.

[vanchutr]

To Ian,
Thank you to answer
I'd followed all instructions of Chooselocale program. You could see these snapshots

[MU]

maybe you need more files like the folder
/usr/lib/locale/UTF-8

Here is a tar.gz to unpack with xarchive.

extract somewhere, then copy to /usr/lib/locale/ what you need.

http://dotpups.de/tests/german-utf8-usr-lib-locale.tgz

contains:

Code: Select all

sh-3.00# ls -1 locale
de
de_AT
de_AT.UTF-8
de_BE
de_BE.UTF-8
de_CH
de_CH.UTF-8
de_DE
de_DE@euro
de_DE.UTF-8
de_LU
de_LU.UTF-8
en_AU
en_US
en_US.UTF-8
ISO-8859-15
UTF-8
sh-3.00# 

I do have problems.
One is concerning rxvt, but I could solve that:
http://www.murga-linux.com/puppy/viewto ... rm&t=13302

the other one:
I use Icewm.
If I create a utf-file like test-äöüß.txt, and open it in leafpad, the titlebar cuts everything in the title after the first utf-character, so it displays only "test-"

/tmp/xerrs.log tells me:
Error converting string 'testäö--ü.txt' to UTF-8 from current locale: Invalid byte sequence in conversion input
Error converting string 'testöäääääää.txt' to UTF-8 from current locale: Invalid byte sequence in conversion input
Error converting string 'tät' to UTF-8 from current locale: Invalid byte sequence in conversion input
Error converting string '1111ä.txt' to UTF-8 from current locale: Invalid byte sequence in conversion input
Error converting string 'testäöü.txt' to UTF-8 from current locale: Invalid byte sequence in conversion input
Error converting string 'testäöü.txt' to UTF-8 from current locale: Invalid byte sequence in conversion input
Error converting string 'testäöü.txt' to UTF-8 from current locale: Invalid byte sequence in conversion input
Warning: Cannot convert string "nil2" to type FontStruct


I also tried icewm 1.2.29, same result.
It seems icewm tries to convert utf to utf, what creates "doubled" utf-sequences, which are illegal of course.
Mark

[MU]

found solution for the icewm-titles I think.
If you use a UTF-8 locale, you need a icewm compiled with
--disable-i18n

I compiled it with
./configure --prefix=/usr/local --enable-gradients --enable-antialiasing --enable-guievents --with-icesound=OSS,Y,ESound --disable-i18n

Can upload a dotpup tomorrow.
Mark

[GuestToo]

do you mean like this?
http://i16.tinypic.com/33u4mpu.png

leafpad ... Icewm 1.2.29

./configure --build=i486 --enable-gradients --enable-antialiasing --enable-guievents --with-icesound=OSS,Y,ESound --with-imlib

[MU]

GuestToo, I'm shure you have no LANG=xx_XX.UTF-8

With such a setting - no go.
Only if I compile with
--disable-i18n , but then there are no translations available.
I tried the whole evening to find a solution in the sourcecode, but my knowledge in C++ is not good enough

Mark

[GuestToo]

does Icewm use iconv ? ... if it does, you could start iconv with a wrapper script to get the parameters ... something like this:

#!/bin/sh
echo "$@" >> /tmp/iconv.log
exec /usr/bin2/iconv "$@"

it might give you a clue to what is happening

there are compile config options for what version of utf Icewm uses internally ... for example, there are messages like these when running ./configure

checking whether iconv converts from UTF-8 to UCS-4//TRANSLIT... yes
checking whether iconv converts from ISO-8859-1 to UCS-4//TRANSLIT... yes
checking whether iconv converts from ISO-8859-2 to UCS-4//TRANSLIT... yes
checking whether iconv converts from UTF-8 to UCS-4//TRANSLIT... yes

[MU]

it uses a inbuild iconv function from libc (#include iconv.h).

It is called from
ylocale.cc, around line 148

Code: Select all

    if (0 > (int) iconv(instance->toUnicode, &inbuf, &inlen, &outbuf, &outlen)){
        warn(_("Invalid multibyte string \"%s\": %s"), lStr, strerror(errno));
    }

You can add different utf-versions a bit before, line 80:

Code: Select all

/// TODO #warning "this is getting way too complicated"

    char const * unicodeCharsets[] = {
#ifdef CONFIG_UNICODE_SET
        CONFIG_UNICODE_SET,
#endif    
//      "WCHAR_T//TRANSLIT",
        (*endian.c ? "UCS-4LE//TRANSLIT" : "UCS-4BE//TRANSLIT"),
//      "WCHAR_T",
        (*endian.c ? "UCS-4LE" : "UCS-4BE"),
        "UCS-4//TRANSLIT",
        "UCS-4",
        NULL
    };

I just tried some other values reported by

iconv -l
but without success.
But I also think, it must be here.
However, I fear the problem is:
iconv seems not to be able, to convert utf to utf.
As a file saved with leafpad using a UTF-8 locale already has utf-characters in the filename, you get this error about illegal Multibyte-characters.

So I also tried to bypass the conversion, but also without success.
But this might be due to my lack in C++ knowledge, maybe I just wrote wrong code.

Also it would not be that simple.
As this function IS needed to read .desktop-files.
Without it, UTF-characters inside such files are displayed wrong.

So I also had tried to modify other places like wmclient.cc, but that confused me even more

Mark

[MU]

ok, the problem was determinating the SOURCE-codeset.
This is taken from the locale-settings.
To open .desktop-files, it must be xx_XX.UTF-8
To display filenames that include utf-characters in the titlebar, it must be codeset = "ISO-8859-1";

At least with german characters, I did not try vietnamese or arab.

My modified versions for ylocale.cc are here:
http://dotpups.de/tests/icewm-1.2.29-unicode/

I added a check:
if there is an error using UTF-8 (this multibyte-error), then it is tried again with "ISO-8859-1".
Now I have Umlaute in the menu AND the titlebars for files saved with leafpad using UTF-characters in the filename.

I also uploaded wmapp.c, where I commented the "restart xterm" menu-entry, as it just kills icewm, but not X.

I will try to upload Muppy006 tomorrow night.
Might not include everything I wanted for this release, but I hope I fixed most of the bugs, and Unicode should now work in most cases (Xfe not really, but can be used partially after another code-modification).
Default filemanager is rox 2.5.

I'm away this weekend, so I'd like to release it tomorrow already.
Mark