Puppy Linux Discussion Forum

[Aitch]

The Origin of this idea is here

[maddox]

Hi Aitch,
I'd love an International Puppy.
I'm not a great programmer (beginner), but BarryK's scripts have very handy comments at strategic places to help those who WANT to learn.

So I've been helping translate Toutou (french Puppy).
Our big probs are UTF-8, localisations and some progs/scripts need ISO-8859-1 encoding for french accents while others need UTF-8 to display french accents correctly.
A simple example : "xmessage" needs iso-8859-1 encoding for french accents while "Xdialog" needs utf-8
and what do you do with a script that contains both ? .. i.e. bootmanager
All binarys need international support compiled in !

Is the kernel international enabled ? The nls_utf_8.ko.gz isn't included in "initrd.gz" nor in Puppy.
The foreign users can't select their correct kbd at Puppy boot time (just typing "puppy pfix=ram" can be hell -> qwerty/qwertz/azerty...)
The specific language xx_xx.uft-8 locale isn't included either, but this is language specific.

On another note, most translators have very limited (non existant) script understanding but are very willing to do the translations.

"esmourguit" uses "po/mo" files that he finds on the net (maybe not the latest version)
But for this to function, the programmers have to do some extra work : use variables in the script and the variables contain the text to translate -> "po" file
They (or the team) are responsible for keeping the "po" file updated for newer versions of the prog !
the translators "translate" the specific language "po" file and submit it for approval.

I probally posted in the wrong topic.. sorry
.. but the heading was exactly what was needed.
.

[MU]

Kernel 2.6.27 uses UTF by default.
I had to modify Muppys init scripts when I upgraded from 2.6.21.7 from iso-code to 2.6.27 with utf8.

You can get a Puppy with that Kernel here:
http://www.murga-linux.com/puppy/viewtopic.php?t=34341

Maybe also Puppys 2.6.25 uses UTF, I did not check that yet.
Xmessage:
replace it with a symlink to gxmessage.

Then you can display UTF like this:
xmessage -encoding UTF-8 "äöü"

Mark

[Aitch]

Thanks maddox, I wasn't aware of this ISO-8859-1 encoding anomaly

Mu, I think your Muppy is probably the best production semi international puppy, & was wondering if the work you have already done with German could be a model for others? Have you experience of any accent anomaly as maddox says?

However, I am aware of the French/Russian/Japanese/Hacao_Vietnamese puppies also, as well as the 41 language Puppy, which uses SCIM or SCIM Bridge to get international characters up

Do you know if all language texts will work with SCIM Bridge?
Is this the common UTF8 solution, for accents & boot keyboard availability??

I believe full internationalisation can be broken down to:-

1 Keyboard selection at boot/(kernel issues?)
2 Character text display in all programs/editors
3 Translations of scripts within Puppy core
4 Translation of Howto / Help files
5 Correct display of browser fonts including flash/browser stability
6 Internationalised Psip/comms progs

Is there anything else? Any other hurdle?

What translation problems still remain, as maybe volunteers can be found, here?

How are Greek, Dutch & Spanish doing? Any others?

See this list

http://www.puppylinux.org/wiki/multi-lingual-puppy/language-chooser/some-localisations

It would be good if people who already have one language characterset/translation working, could maybe try to see
if their working version can be modified to co-exist with another language

e.g. Muppy + Toutou
Russian + German
Toutou + Russian
Vietnamese + Japanese

or something similar/crossed over....?

I realise this is a big leap, and many people are just happy too get Puppy working at all, however, anyone feels like experimenting......?

Please post any results/anomalies/problems, and let's see if this is possible

Raffy, Magerlap, Nyu, Irihapeti, Béèm, anyone else, [sorry if I missed you] any ideas/input?

Lobster, can you assist? Is this worth a new Wiki page, yet? ....or maybe just add something here?

http://www.puppylinux.org/wiki/multi-lingual-puppy

[I see this as a step towards St Bernard /Rescue Pup, for anyone interested...]

MU's excellent 76 db Dictionary, for people not aware of it,
... can this be made into an SFS? For incorporation, Mark?

http://www.murga-linux.com/puppy/viewtopic.php?t=7976


This may be a leap of faith, but it's Xmas, so GoodWill to all Good [International] Puppians [do bad ones really exist?]

Here's one I like..



Courtesy: http://www.news.com.au/story/0,23599,22050638-2,00.html

......Well, someone has to push the boat on this one, & now I did


Aitch

[magerlab]

there's a bunch of material that russian puppy team compiled on russification

we even had to walk through almost the whole linux-in-russian-problems ourselves
and have almost 100% russian puppy( puppyrus) based on puppy 3-4

but the first thing is
UTF-8 must be used!!!!
and surprisingly when i booted new unnamed pupplet(made by wow) with ....27 kernel
i saw unicode locales
also it's the first puppy where usb-pen drive with russian filenames with vfat fs( from windoze) is mounted correctly in pcmanfm

the second thing is to make all puppy scripts support gettext
unless we need to rewrite them

[magerlab]

i forgot...
we can start with having menus and all the menu entires in different languages
that's very easy to do but it needs to be gathered in one place and supervised

i made a basic rusiian pack for puppy 4.1 ( only for gui, not scripts)

and i think also that it's better to have different lang packs than having all the stuff in one
( but menu can be shared through packs)

[amigo]

Using gettext or xgettext in the scripts is the only reasonable thing to do. That way, the text to be translated is not completely mixed in with the actual code. This would mean using bash as the script language as it can work with LC variables.
This is very important and critical because most of what makes puppy Puppy, is done with scripts. Still, there are probably many of the commonly-used Puppy binary programs which lack internationalization. Working on adding i18n support to these programs would make a good project for anyone who wants to get their feet wet with C and/or C++ coding, without having to first become a master. Adding the i18n support is not usually too difficult and is completely separate from doing the actual translations. The same is true for using gettext in scripts -adding support for gettext is a separate task from actually translating.

Here's a link to a basic bash/i18n howto:
http://www.cs.ualberta.ca/doc/UNIX/BASH_Guide/localization.html

[Aitch]

[maddox]

The "petget manager" and the pet's displayed in the download list needs to be International too (no language barrier if at all possible).
A user might be multi-lingual, and prefer to install a different language pet instead of using the current locale ..
i.e: a french user, but they prefer to install certain pet's in english or another favorite language...
.. This implies a strict compliance on how to create/translate a pet and use the correct encoding for the desired language,
and opening up the petget manager to translated pet's.

We've translated quite a number of "pet's" , so how do we integrate these language/version specific pets ?
A language specific site/link to download pet's might be an idea.

[straypup]

Sorry for such a late post, but has anyone checked out the latest Firefox. Just downloaded the latest version in Thai and installed on 4.21 Final, also installed Thai unicode fonts, and SCIM (SCIM-bridge method) for my wife. Everything seems to be working, Firefox even selects native Thai sites. Just an idea of how to get basic internationalization started.
stray

[Aitch]

bump

Just found this very useful link

http://roscidus.com/desktop/Internationalisation

[L18L]