RFR: JDK9 message drop interim L10n resource update

Robert Field robert.field at oracle.com
Thu Nov 17 17:04:53 UTC 2016


On 11/17/16 08:29, Leo Jiang wrote:
> Hi Shinya, Robert,
>
> 1.
> After review these messages, I found the problem was caused by the 
> word order of original English sentence.
>
> For the below messages, they have same pattern, the type is the suffix 
> after colon.
> > jshell> int n = 0
> > n ==> 0
> > |  次を作成しました: 変数 n : int
> >
> > jshell> n = 0
> > n ==> 0
> > |  nに割り当てられました : int
> >
> > jshell> n
> > n ==> 0
> > |  nの値: int
>
> @Robert,
> Would you help to confirm if we need to keep the {name} : {type} 
> together? Currently translator might think the ': {type}' must be in 
> the end of sentence, and can be separated from {name}.
>
> The English message:
> '{result}{pre}assigned to {name} : {type}{post}'

Well, the problem is I mixed English with a made-up syntactic language 
and that made-up language doesn't work with a natural language with a 
different word order.

The ":" means "of type" and they (probably) need to be together. So, 
"assigned to g : int", means "assigned to g of type int".

This is for verbose mode, so what might be best is to just spell out "of 
type".  The problem is that for readability the type is best at the end, 
as it can be complex:

     jshell> Map<Integer, Map<String, StringBuilder>> map = new HashMap<>()
     map ==> {}
     |  created variable map : Map<Integer, Map<String, StringBuilder>>

What would work well in Japanese and Chinese?

Thanks,
Robert

>
> 2.
> @Shinya, Thank you for providing us patch, it's so kindly. Bug I can't 
> use your patch as the patch was created according to the webrev which 
> specially been converted for human reading. Would you edit a English 
> resource file with the correct Japanese lines (only several lines you 
> want to fix), and send the result to me via email? Thank you, very 
> appreciated!
>
> Thanks,
> Leo
>
> On 11/16/2016 12:47 PM, ShinyaYoshida wrote:
>> Hi Leo,
>> Thank you for updating translation.
>> In jshell, the most becomes better than previous one!
>>
>> I've noticed incorrect word order in jshell translation:
>> jshell> /set feedback verbose
>> |  フィードバック・モード: verbose
>>
>> jshell> int n = 0
>> n ==> 0
>> |  次を作成しました: 変数 n : int
>>
>> jshell> n = 0
>> n ==> 0
>> |  nに割り当てられました : int
>>
>> jshell> n
>> n ==> 0
>> |  nの値: int
>>
>> jshell> void m(C c) {}
>> |  次を作成しました: メソッド m(C)まで、参照できません。しかし、 
>> class Cが宣言される
>>
>> (1): The type name goes far from the variable name.
>> (2): "まで、参照できません。しかし、 class Cが宣言される" has 
>> incorrect order.
>>
>> They should be following:
>> jshell> /set feedback verbose
>> |  フィードバック・モード: verbose
>>
>> jshell> int n = 0
>> n ==> 0
>> |  次を作成しました: 変数 n : int
>>
>> jshell> n = 0
>> n ==> 0
>> |  n : intに割り当てられました
>>
>> jshell> n
>> n ==> 0
>> |  n : intの値
>>
>> jshell> void m(C c) {}
>> |  次を作成しました: メソッド m(C)。しかし、 class Cが宣言されるま 
>> で、参照できません
>>
>> Here is fix: 
>> http://cr.openjdk.java.net/~shinyafox/kulla/8169618/l10n_ja.properties.patch.wdiff
>> This is a wdiff against your udiff at the last 3 lines of
>> langtools/src/jdk.jshell/share/classes/jdk/internal/jshell/tool/resources/l10n_ja.properties. 
>>
>>
>> For maintainability, I'd like to suggest breaking lines like 
>> startup.feedback does so
>> in jdk/internal/jshell/tool/resources/l10n.properties.
>>
>> Regards,
>> shinyafox(Shinya Yoshida)
>>
>>
>> 2016-11-14 15:09 GMT+09:00 Leo Jiang <li.jiang at oracle.com 
>> <mailto:li.jiang at oracle.com>>:
>>
>>     Hi,
>>
>>     Please review:
>>     Webrev:
>>     http://cr.openjdk.java.net/~ljiang/8169618/webrev/read/ 
>> <http://cr.openjdk.java.net/~ljiang/8169618/webrev/read/>
>>
>>     for bug:
>>     https://bugs.openjdk.java.net/browse/JDK-8169618 
>> <https://bugs.openjdk.java.net/browse/JDK-8169618>
>>
>>     Please help us to review
>>      - over-translate, e.g. keywords should be not translated
>>      - not-translate, e.g. some sentences or strings are not 
>> translated while they should be.
>>      - any programming syntax error
>>      - multiple lines property with correct line ending '\n\'
>>      - the position of placefolder after translation
>>      - any other issues.
>>
>>     Thanks,
>>     Leo
>>
>>



More information about the jdk9-dev mailing list