[icedtea-web] resurrection of (output)console?

[Jiri Vanek]

On 11/06/2013 01:49 PM, Omair Majid wrote:
> * Jiri Vanek <jvanek at redhat.com> [2013-11-06 03:31]:
>> On 11/05/2013 08:44 PM, Omair Majid wrote:
>>> * Jiri Vanek<jvanek at redhat.com>  [2013-11-04 06:31]:
>>>> before you will wonder about removed redirectStreams, please test
>>>> patch :). Also it is remove in 1.5. Imho "debug" should have nothing
>>>> to do with "redirecting" also it was used wrongly, and made consoel
>>>> even less usefull...
>>>
>>> Let me break down the code:
>>>
>>> If the user has asked for logging, then log to java.stdout and
>>> java.stderr as well as the actual stdout/stderr.
>>>
>>> If the user has said ICEDTEAPLUGIN_DEBUG=true, then the C side of the
>>> plugin will be printing it's information on stdout/stderr. In that case,
>>> redirect the java-side to java.stdout and java.stderr.
>>>
>>> Can you explain the problem with this scheme?
>
> Thanks for explaining! There's some things which I agree are a bug and
> should be fixed. Some other things, I am not sure about.
>
>> What is wrong, is that debug is suddenly changing not the verbosity of
>> output, but destination of output(without changing the verbosity
>> actually).
>>
>>   - In 1.4 we have javaconsole and stdout/err as debugging outputs
>>
>>   - there is variable redirectStreams which is true if
>>   ICEDTEAPLUGIN_DEBUG is declared
>>
>>     - in 1.4 ICEDTEAPLUGIN_DEBUG was affecting only C part, so it is
>>     suspicious
>>
>>   - there is variable enableLogging, which is always false, because it
>>   is read from properties, *before* they are read from file (bug in
>>   1.4, fixed in head) (it also depneds on -verbose, but for simplicity
>>   i will stay of as this (console) is just for applets part in 1.4)
>>
>>     - as it is used, it is also duplicated withs console
>>     DISABLED/all_other_values
>>
>>   - now the original "logging" was under condition redirectStreams ||
>>   enableLogging, so it was depnding only on value of redirectStreams
>>   and so on ICEDTEAPLUGIN_DEBUG set.
>>
>>    - suspicions isnt it?
>
> No, not if you look at the actual effects.

ok. With your latest sentence this gave a bit more sense.

>
>>   - if one declared ICEDTEAPLUGIN_DEBUG (whic is normally affecting verbosity of C part), sudddenly
>>     -  the console starts to work
>
> Okay, this is a bug that we should fix.
>
>>     -  the verbosity of java part is still not-debug (so nothing interesting in console)
>
> That should be controlled by 'deployment.trace' option.

which was never implemented (and got removed in 1.5 .. so not worthy to fix)
>
>>     -  the stdou/err is suddenly silenced
>
> No, the C side of the plugin is using the stdout/stderr of the browser.
> It will print tons of information. The java side writes its
> stdout/stderr to another location to keep things separate.

oooooh.....yes. Now it gives much more sense. That you wish to keep the logs separated is something 
what never crossed my mind. (I still doubt its  useful, but at least I see your point!)

Ok. With  " C side of the plugin is using the stdout/stderr of the browser.
 > It will print tons of information. The java side writes its
 > stdout/stderr to another location to keep things separate" I think we can close this issue and I 
will keep importance of separated logs for 1.5 in mind.

J.