[rfc][icedtea-web] refactored logging

[Jiri Vanek]

On 10/22/2013 04:06 PM, Omair Majid wrote:
> * Jiri Vanek<jvanek at redhat.com>  [2013-10-22 09:58]:
>> On 10/22/2013 03:42 PM, Omair Majid wrote:
>>> +    /** For testing only: throw away the previous config */
>>> +    static void resetLogConfig() {
>>> +        if (logConfig != null) {
>>> +            logConfig = new LogConfig();
>>> +        }
>>> +    }
>>> +
>>>       public String getIcedteaLogDir() {
>>>           return icedteaLogDir;
>>
>> Well, much better :) But, isn't set/get much more common?
>
> It probably is. But I don't think it makes sense here. We don't want to
> allow general code to set a new LogConfig, do we?
no - for this, the package private or reflection should be enough.

It's only for tests.
> And when I look at the tests, the only thing they care about is
> resetting it to the default value.

Yes, thats what I'm worying a bit. That you do not do do backup, restore, but just reset. However it should be ok here..
>
> Do you expect any code to ever subclass a LogConfig and use that?
> setLogConfig may be worth having in that case. Otherwise, you can just
> call set* on individual settings of a LogConfig.
>
>> Again, do as you wish :) I'm ok with all three variants.
>
> Okay, I will push this patch shortly (unless you (or anyone else) have
> any concerns).
>
> Cheers,
> Omair