RFR(JDK 14/java.xml) 8223291: Whitespace is added to CDATA tags when using OutputKeys.INDENT to format XML

Daniel Fuchs daniel.fuchs at oracle.com
Tue Jul 2 16:16:08 UTC 2019


Hi Joe,

On 02/07/2019 17:51, Joe Wang wrote:
> Thanks Daniel. That test case is added.

LGTM!

best regards,

-- daniel

> 
> Best,
> Joe
> 
> On 7/2/19, 1:21 AM, Daniel Fuchs wrote:
>> Hi Joe,
>>
>> I haven't spotted anything obviously wrong.
>>
>>> The content of "abc<![CDATA[ & ]]>xyz" should be "abc & xyz", not 
>>> "    abc\n & \n    xyz\n" as that before this fix. 
>>
>> Maybe the test could have a test case for that specific example too?
>> Just to make sure whitespaces in CDATA aren't eaten away?
>>
>> best regards,
>>
>> -- daniel
>>
>>
>> On 01/07/2019 20:46, Joe Wang wrote:
>>> Please review a fix to xml pretty print. This is a regression 
>>> introduced during the JDK 9 development. CDATA is marked up to be 
>>> interpreted literally as textual data, in other words, it's still 
>>> character data. The processor therefore shall treat it as text data. 
>>> The content of "abc<![CDATA[ & ]]>xyz" should be "abc & xyz", not 
>>> "    abc\n & \n xyz\n" as that before this fix.
>>>
>>> JBS: https://bugs.openjdk.java.net/browse/JDK-8223291
>>> webrev: http://cr.openjdk.java.net/~joehw/jdk14/8223291/webrev/
>>>
>>> Thanks,
>>> Joe
>>>
>>



More information about the core-libs-dev mailing list