[rfc][icedtea-web] refactored logging
jvanek at redhat.com
Wed Sep 25 09:52:58 PDT 2013
On 09/24/2013 11:57 PM, Omair Majid wrote:
> On 09/24/2013 01:12 PM, Jiri Vanek wrote:
>> On 09/24/2013 04:31 PM, Omair Majid wrote:
>>> Hi Jiri,
>>> On 09/24/2013 02:43 AM, Jiri Vanek wrote:
>>>> One possible improvement (related to this change set) :
>>>> There is 138 occurrences of log(Level level, Throwable s)
>>>> There is 67 occurrences of log(Throwable s)
>>>> The log(Throwable s) is call to log(Level.ERROR_DEBUG, (Object) s);
>>>> and those log(Level level, Throwable s) are mostly
>>>> log(Level.ERROR_ALL, throwable) calls
>>>> I was thinking about changing it to
>>>> There is 68 occurrences of log(Level level, Throwable s)
>>>> There is 138 occurrences of log(Throwable s)
>>>> where The log(Throwable s) will be call to log(Level.ERROR_ALL,
>>>> (Object) s)
>>>> What do you think?
>>> I think this is one of those decisions where having some data would
>>> allow for a better evaluation. Is the code calling log(Throwable)
>>> expecting it to be user-facing or merely debugging information ?
>> I believe all those exceptions are moreover deadly, rare or serious..
>> Not sure if this helps:)
> Then lets make it ERROR_ALL. I reserve the right to ask for this to be
> reverted after seeing some logs :)
At the end I kept it as it was. I think it should be statement to log always. And both err/out log
should have same api.
Few changes on last second:
fileloger have now append possibility and close possibility
the message consumer is now dameon.
good tests :) Especially I was happy teh paralel logger test passed always :)
Thank you very much for review!
More information about the distro-pkg-dev