[rfc][icedtea-web] mov repeated trusAll/none/shouldPromt/isHeadless to one palce

[Andrew Azores]

On 17/06/15 05:45 AM, Jiri Vanek wrote:
> On 06/16/2015 05:32 PM, Jiri Vanek wrote:
>> Unluckily the test needed a lot of work, because now the message is 
>> always posted to queue.
>>
>>
>> Otherwise the patch is pretty straightworward.
>>
>> Aslo I have encountered some issues with reprodcuers:
>>
>> FAILED: testCaching(AbsolutePathsAndQueryStrings) 
>> expected:<...ipHttpPathParams.jar[]> but
>> was:<...ipHttpPathParams.jar[.i_abcd]>
>>
>>     - query patch is guilty
>>
>> Those three are very suspicious.
>> FAILED: ReplaceSecurityManagerLunch1(ReplaceSecurityManagerTest) 
>> ReplaceSecurityManagerLunch1 should
>> not be terminated, but was
>> FAILED: 
>> ApplicationJNLPLocalTest(CodeBaseManifestEntrySignedNotMatching) 
>> should contain `*** APPLET
>> FINISHED ***`
>> FAILED: AppletJNLPRLocalTest(CodeBaseManifestEntrySignedNotMatching) 
>> should contain `*** APPLET
>> FINISHED ***`
>>
>> those three seems to be ths patch related....hmmm
>> FAILED: testJavawsJNLP - opera(PackGZipTest) stdout should contain 
>> init, but it didnt.
>> FAILED: testJavawsJNLP - midori(PackGZipTest) stdout should contain 
>> init, but it didnt.
>> FAILED: testJavawsJNLP - epiphany(PackGZipTest) stdout should contain 
>> init, but it didnt.
>>
>>
>> I needed to temrinate testun so I have not run them all. But seem 
>> sthat soemthing already went wrong
>> in alst few pushes. (will run them today overnight)
>>
>> J.
>>
>>
>>
> Looks like I overlooked
>
>   private void checkTrustWithUser() throws LaunchException {
>         if (JNLPRuntime.isTrustNone()) {
>             if (!securityDelegate.getRunInSandbox()) {
>                 setRunInSandbox();
>             }
>             return;
>         }
>         if (JNLPRuntime.isTrustAll() || 
> securityDelegate.getRunInSandbox()) {
>             return;
>         }
>
>         if (getSigningState() == SigningState.FULL && 
> jcv.isFullySigned() && !jcv.getAlreadyTrustPublisher()) {
>             jcv.checkTrustWithUser(securityDelegate, file);
>         }
>     }
>
>
> =>
>
>   private void checkTrustWithUser() throws LaunchException {
>
>
>         if (getSigningState() == SigningState.FULL && 
> jcv.isFullySigned() && !jcv.getAlreadyTrustPublisher()) {
>             jcv.checkTrustWithUser(securityDelegate, file);
>         }
>     }
>
>
>
> ok?
>
>
>

I haven't run tests on this but it looks pretty good to me, definitely 
nice to have this security decision logic a bit more centralized.

-- 
Thanks,

Andrew Azores