Hi all!
I'm working on localization of OpenOffice.org 2.0 in Armenian.
More than one year ago we (Open Source Armenia team) have localized
OpenOffice.org 1.1.0 but because of some problems we couldn't continue our
work. But now we've already started localization of OpenOffice.org 2.0. I
hope that the design of http://hy.openoffice.org will be completed in few
days, and will be able to show you all the works done.
During localization of v1.1, I've made some changes without discussing them
in mail list.
Now I want you to introduce some solutions which are determined by our
language specification.
----------------------------------------------------------------------------
In punctuation we use '.' and ':' but in Armenian they have opposite
meanings.
'.' (Dot) in Armenian is the same as ':' in English.
':' (Colon) in Armenian is the full stop symbol.
We don't use the '!','?' symbols.
My changes in svx/source/editing/svxacorr.cxx
SvxAutoCorrect::FnCptlSttSntnc
....
bool is_armenian = false;
if (eLang == LANGUAGE_ARMENIAN)
is_armenian = true;
switch (*pStr)
{
// Armenian full stop
case ':' :
case 0x0589:
{
if( (nFlag & C_FULL_STOP) || (!is_armenian) )
return FALSE;
..
Of course these are not the all changes.
And in file sw/source/core/edit/autofmt.cxx
SwAutoFormat::AutoCorrect
.....
case '.':
case '!':
case '?':
if( (aFlags.bCptlSttSntnc) && (eCharClassLang != LANGUAGE_ARMENIAN) )
bFirstSent = TRUE;
case ':':
if( (aFlags.bCptlSttSntnc) && (eCharClassLang == LANGUAGE_ARMENIAN) )
bFirstSent = TRUE;
.....
These changes made to correct automatic uppercasing of the first letter
after '.', '!', '?' symbols in Armenian texts and resetting ':' symbol as
full stop.
--------------------------------------------------------------------------
One question about uppercasing:
Uppercasing of the symbol 0x0587 gives two symbols: 0x0565, 0x0582.
This rule described in MACHINE-GENERATED file:
Icu/wntmsci10.pro/misc/build/icu/source/common/uniset.cpp
CASE_NONPAIRS[]
That's not exactly quit right!
This type uppercasing is right in western Armenian which is now used among
Armenian Diaspora, but in official orthography the uppercase of that symbol
is the union of 0x0535 and 0x054e.
In MS Office they are just ignoring uppercasing of this symbol.
So how can we decide this issue?
--------------------------------------------------------------------------
MySpell
There isn't any 8-bit encoding in Armenian. Windows supports only Unicode
encoding for Armenian. And as you know MySpell supports only ANSI so that's
why we've decided to add one pseudo encoding. Something similar is used in
Armenia for Windows users. There are 2-3 widespread programs which are used
as text write tool. But each of those programs uses its own encoding. So to
exclude any conflict and problem with encoding we've decided to refuse using
Armenian 8-bit encoding in OpenOffice.
File: lingucomponent/source/spellcheck/spell/sspellimp.cxx
Method: SpellChecker::GetSpellFailure
rtl_getTextEncodingFromUnixCharset sets encoding to RTL_TEXTENCODING_UTF8.
In MySpell I've added 2 methods: encoder (UTF-8 -> ARMSCII8) and decoder
(ARMSCII -> UTF-8).
File: lingucomponent/source/spellcheck/myspell/myspell.cxx
Method: MySpell::cleanword
I've described another special_chars for Armenian language.
----------------------------------------------------------------------------
translate-toolkit
localize tool exports en-US.SDF file successfully.
But oo2po tool in some files makes wrong exports.
As an example:
File: basic\source\ app.pot
#: basic.src#IDD_TT_ABOUT_DIALOG.1.fixedtext.text
msgid "_: basic.src#IDD_TT_ABOUT_DIALOG.1.fixedtext.text\n"
"VCLTestTool"
msgstr ""
How can I fix this problem?
----------------------------------------------------------------------------
In localization of v.1.1 we've used language code 98.
But now as I see there isn't any requirement for it, is it?
----------------------------------------------------------------------------
In the article at
http://l10n.openoffice.org/L10N_Framework/ooo20/localization_of_openoffice_2
.0.html there are described new localization rules. Is this mean that now
localization can be done without direct using CVS, and all patches will be
applied to CVS by OpenOffice developers?
Thanks!
Best regards!
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
