Launchpad itself

Use appropriate marking for KDE plural forms

Bug #74361 reported by Данило Шеган on 2006-12-04

Affects		Status	Importance	Assigned to	Milestone
	Launchpad itself	Won't Fix	Low	Unassigned

Bug Description

Inside kdelibs, there's a message which is replacement for regular gettext "Plural-Forms" header.

Specific message is:

"_: Dear translator, please do not translate this string in any form, but pick the _right_ value out of NoPlural/TwoForms/French... If not sure what to do mail <email address hidden> and <email address hidden>, they will tell you. Better leave that out if unsure, the programs will crash!!
Definition of PluralForm - to be set by the translator of
kdelibs.po"

(i.e. https://translations.launchpad.net/distros/ubuntu/edgy/+source/kdelibs/+pots/kdelibs/sr/1005/+translate at the moment)

The translation for that should be limited to one of:
NoPlural: 0
TwoForms: n != 1
French: n > 1
OneTwoRest (???):
Russian: n%10==1 && n%100!=11 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2
Polish: n==1 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2
Slovenian: n%100==1 ? 0 : n%100==2 ? 1 : n%100==3 || n%100==4 ? 2 : 3
Lithuanian: n%10==1 && n%100!=11 ? 0 : n%10>=2 && (n%100<10 || n%100>=20) ? 1 : 2
Czech: n%10==1 && n%100!=11 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2
Slovak: n%10==1 && n%100!=11 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2
Arabic: n == 1 ? 0 : n == 2 ? 1 : 2
Balcan:
Macedonian: (n % 10 == 1 && n % 100 != 11) ? 0 : 1
Gaeilge (???): n < 2 ? 0 : n == 2 ? 1 : 2
Maltese: n == 1 ? 0 : (n == 0 || ((n % 100) >= 2 && (n % 100) <= 10) ) ? 1 : ((n % 100) >= 11 && (n % 100) <= 19 ) ? 2 : 3

(left is the KDE value to be used as translation for above message, right is our current definition inside language table; where values are missing, I'll fill them in later)

My idea is to extend language table in Rosetta using left-hand values and use that on PO export.

Blocked by bug #46982.

See original description

Tags:

Данило Шеган (danilo) on 2006-12-04

Changed in rosetta:
status:	Unconfirmed → Confirmed
description:	updated

Revision history for this message

Carlos Perelló Marín (carlos) wrote on 2006-12-15:

Does it mean that they only support that fixed list of plural forms?

I guess the plural form expression is compatible with what we have in gettext, right? (at least the simple ones are quite similar or equal).

Oh, I just saw the Kurdish solution. This plural form implementation is even more broken than I thought. At least, KDE is moving now to standard plural forms....

Revision history for this message

Данило Шеган (danilo) wrote on 2006-12-15:

It's not really broken, the design is same as for Rosetta: system-wide definitions of plural forms. It's just not as extensible as GNU gettext stuff is (which can sometimes hurt us as well).

And GNU gettext reasoning behind not hard-coding plural-forms definitions was that they don't know them all yet, and code changes are slow to reach the masses (the problem we don't have with Rosetta; not sure how KDE handles this, but they probably just release a new kdelibs package when needed).

Today, we're mostly familiar with all possible plural forms definitions, so we can as well hard-code them.

Revision history for this message

Erdal Ronahi (erdalronahi) wrote on 2006-12-15:

I didn't understand Carlos' remark about the "Kurdish solution". Please explain.

Revision history for this message

Carlos Perelló Marín (carlos) wrote on 2006-12-15:

"Kurdish solution" == what Kurdish team (in fact, you did it) added to kdelibs to be able to use plural forms. I was not aware of the way KDE selected the way plural forms work. The only thing I was aware of is the way the tag plural forms.

You can see what I mean at:

https://translations.launchpad.net/distros/ubuntu/edgy/+source/kdelibs/+pots/kdelibs/ku/1005/+translate