/hg/icedtea-web: Introduced logging bottleneck

[Omair Majid]

On 10/01/2013 02:07 AM, Jiri Vanek wrote:
> On 09/30/2013 09:01 PM, Andrew Azores wrote:
>> On 09/25/2013 12:50 PM, jvanek at icedtea.classpath.org wrote:
>>> changeset a817bb6d12a6 in /hg/icedtea-web
>>> details:
>>> http://icedtea.classpath.org/hg/icedtea-web?cmd=changeset;node=a817bb6d12a6
>>>
>>> author: Jiri Vanek <jvanek at redhat.com>
>>> date: Wed Sep 25 18:50:18 2013 +0200
>>>
>>> Introduced logging bottleneck
>>>
>> As noted by Omair and I in the IRC channel today, this patch makes
>> "make check" rather noisy. I also mentioned that a fair number of
>> reproducers are broken by this. The common trait to most/all of them
>> is that they expect stdout to be a particular (very short, common case
>> being <= 2 chars) length, which it is because messages are now coming
>> through stdout which were previously on stderr. For example,
>> CreateClassLoaderTest.CreateClassLoaderLunch1 fails because "netx:
>> Launch Error: Could not launch JNLP file. ( (access denied
>> ("java.lang.RuntimePermission" "createClassLoader"))) " is printed to
>> stdout instead, and the test includes a check that pr.stdout has
>> length <= 2.
> 
> Both those failures are cuased by "enable logging" == true (by other
> words works for me).  However Omiar reported hat he have cleaned the
> configuration.
> 
> Loging enabled == environment variable || -verbose || deployment property
> 

Okay, but I did not change the configuration (whatever it was) before
the patch vs after. Yet, the output was not so noisy before; it is now.
Given the only change is this patch, something is not quite right with it.

Is there anything I can do to narrow down the cause?

Thanks,
Omair

-- 
PGP Key: 66484681 (http://pgp.mit.edu/)
Fingerprint = F072 555B 0A17 3957 4E95  0056 F286 F14F 6648 4681