[Libreoffice] [PUSHED-3-4][PUSHED[3-5][REVIEW] 3-5, 3-5-0, 3-4 corrupted Indic characters and crash under Windows fdo#44208, fdo#45107

[Caolán McNamara]

On Thu, 2012-01-26 at 18:58 +0100, Christophe Strobbe wrote:
> If there are differences between the zh-hk, zh-mo 
> and zh-tw localisations, is this due to different 
> translators or to differences between computing 
> terminology in these three geographic areas? 

Generally we basically lack a concept of supporting language tags of
e.g. zh-Hant and zh-Hans to describe respectively Traditional and
Simplified Chinese regardless of the territory (i.e. zh-Hans-CN,
zh-Hans-SG).

All we really support at the moment is language_TERRITORY, so for the
specific case of default fonts for a language we have to list all
language_TERRITORY variants when we have a language written in multiple
scripts where different default fonts are indicated for each script.

We more or less get away with it for Chinese as within each territory
only one of the writing systems is in use. There's a similar problem for
Serbian where we ideally would have sr-Cyrl and sr-Latn for Cyrillic and
Latin alphabets. Hack of language_TERRITORY only goes so far given the
"synchronic digraphia" of the two writing systems of equal status within
a single territory. We use "sh" as an ugly hack to fake "sr-Latn" there.

That said, for the Chinese case, while the ideal situation would be a
generic zh-Hant font list to use as defaults for all locales that use
Traditional Chinese, it might still be that zh-Hant-HK might prefer a
different default font zh-Hant, given the Hong Kong Supplementary
Character Set. Maybe not, just speculation.

Anyway, we're drifting a bit off topic I think, hopefully erack will get
a chance at some stage to try and thread nice, e.g. bcp47, language tag
support through LibreOffice to smooth out these kinks.

C.