[rfc][icedtea-web] fixing typo in trusted-only reproducer

[Andrew Azores]

On 27/07/15 01:31 PM, Andrew Azores wrote:
> On 27/07/15 11:08 AM, Jiri Vanek wrote:
>>
>> -Trusted-only: true
>> +Trusted-Only: true
>>
>> in manifest.. so it was actually testing nothing :-/
>>
>> its here since begining 
>> http://icedtea.classpath.org/hg/icedtea-web/annotate/dee18160ba9c/tests/reproducers/custom/TrustedOnlyAttribute/srcs/MANIFEST.MF
>>
>> and more surprising is, that this typo do not affect tests results...
>>
>> I'm in progress of writing new set of tests to trusted-only so imho 
>> no deeper investigatins needed.
>>
>> But maybe original author have something to say:)
>>
>>
>> J,
>
> What do you mean, "it was testing nothing"? I think you're missing 
> that Manifest Attributes are case-insensitive [0][1]. Try changing the 
> case of the attribute in that manifest to TRUSTED-ONLY, to 
> Trusted-Only, and to Trusted-only and notice how the test results are 
> (or should be; they are for me) always the same. Then try removing the 
> attribute entirely and note the test failures.
>
> [0] 
> https://docs.oracle.com/javase/7/docs/api/java/util/jar/Attributes.Name.html
> [1] 
> http://icedtea.classpath.org/hg/icedtea-web/file/dee18160ba9c/netx/net/sourceforge/jnlp/JNLPFile.java#l1045
>

Oh, but regardless of the case-insensitivity of the manifest attribute, 
+1 on pushing this patch correcting the attribute to reflect the 
officially documented spelling of "Trusted-Only" anyway. There 
absolutely should not be any functional impact to this AFAIK but it may 
as well match what the source [0] says anyway.

[0] 
https://docs.oracle.com/javase/7/docs/technotes/guides/jweb/security/manifest.html

-- 
Thanks,

Andrew Azores