W3C home > Mailing lists > Public > public-webapps@w3.org > July to September 2010

Re: [widgets] Buggy P&C i18n Tests

From: Marcos Caceres <marcosc@opera.com>
Date: Mon, 06 Sep 2010 15:05:38 +0200
Message-ID: <4C84E722.4060209@opera.com>
To: Scott Wilson <scott.bradley.wilson@gmail.com>
CC: public-webapps <public-webapps@w3.org>


On 9/6/10 3:03 PM, Scott Wilson wrote:
> Something else went strange with the tests - they have odd new names
> and url's that don't work:
>
> E.g.
>
> Test "b5" has been renamed to:
>
> http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/ta-RRZxvvTFHx/000/ta-rr-000.wgt
>
>  However the actual test widget is still at:
>
> http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/ta-RRZxvvTFHx/000/b5.wgt
>
>  Well, its clearly incorrect. However I would prefer not to have to
> go and manually retype all the URLs for my tests so if we can have
> the old names back that would be nice!

D'oh, our test generator went loopy! :( I'm on it!

> On 6 Sep 2010, at 12:54, Marcos Caceres wrote:
>
>>
>>
>> On 9/6/10 12:25 PM, Scott Wilson wrote:
>>> To be honest I've been having problems with the i18n test cases
>>> generally. For example:
>>>
>>>
>>> Tests that LRO direction applies to the name element. To pass,
>>> the displayed value must render as "םפללחק".
>>>
>>>
>>> I can't get the text in that config.xml to show as anything
>>> other than "קחללפם" - even manually copying it into a HTML file
>>> with the div attribute in every combination and opening in every
>>> browser I've got (including Opera ;-) has the same result!
>>
>> Yeah, we are working on it too. This is by no means easy. It's why
>> the i18n guys made us do it.
>>
>> Try:
>>
>> <p><span style="unicode-bidi: bidi-override; direction:ltr">
>> "קחללפם"</span></p>
>>
>> Should work fine in all browsers.
>>
>>> On 5 Sep 2010, at 23:24, Marcos Caceres wrote:
>>>
>>>>
>>>>
>>>> On 9/4/10 4:54 PM, Scott Wilson wrote:
>>>>> I'm also getting a Zip file error on the i18n-039&   040
>>>>> sets (these are the only ones I've implemented so far):
>>>>>
>>>>> java.util.zip.ZipException: oversubscribed dynamic bit
>>>>> lengths tree
>>>>>
>>>>> They seemed to work OK earlier on yesterday, but seem to
>>>>> have become recently corrupted. (They won't unzip using
>>>>> standard zip desktop tools either)
>>>>
>>>> Re-uploaded all the tests, and we fixed the borked tests too.
>>>> Happy testing :)
>>>>
>>>>> S
>>>>>
>>>>> On 3 Sep 2010, at 19:55, Marcos Caceres wrote:
>>>>>
>>>>>> Hi, Please be warned that following tests have some bugs.
>>>>>> We will fix them next week.
>>>>>>
>>>>>> http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/i18n-lro/027/
>>>>>>
>>>>>>
>>
>>>>>>
http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/i18n-rlo/027/
>>>>>> http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/i18n-ltr/027/
>>>>>>
>>>>>>
>>
>>>>>>
http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/i18n-rtl/027/
>>>>>> http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/i18n-lro/028/
>>>>>>
>>>>>>
>>
>>>>>>
http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/i18n-rlo/028/
>>>>>> http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/i18n-ltr/028/
>>>>>>
>>>>>>
>>
>>>>>>
http://dev.w3.org/2006/waf/widgets/test-suite/test-cases/i18n-rtl/028/
>>>>>>
>>>>>> Effectively, the UA would reject the value of each because
>>>>>> it is not a valid media type, so it would never be
>>>>>> "displayed".
>>>>>>
>>>>>> Also, "If the src attribute of the content element is
>>>>>> absent or an empty string, then the user agent must ignore
>>>>>> this element". This means that the tests above would not
>>>>>> run at all because the UA would reject the content element
>>>>>> (as they don't have a src attribute).
>>>>>>
>>>>>> -- Marcos Caceres Opera Software
>>>>>>
>>>>>
>>>>
>>>> -- Marcos Caceres Opera Software
>>>
>>
>> -- Marcos Caceres Opera Software
>

-- 
Marcos Caceres
Opera Software
Received on Monday, 6 September 2010 13:06:17 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:49:40 GMT