Re: trigger errors

2011-11-16 Thread Ivan Habunek
On 16 November 2011 08:27, Florian Semm wrote: > the coverage of the LoggerConfigurator is now at 90%. May I commit my tests? > > btw: there was no check if a the specified rendererClass/renderingClass > exists. I fixed this in my tests too. That's great. I don't think you have commit privileges

Re: trigger errors

2011-11-15 Thread Florian Semm
Am 14.11.2011 14:53, schrieb Ivan Habunek: On 13 November 2011 14:39, Florian Semm wrote: the configurator and the adapters are good examples for separation of concerns. :) the configurator-test only has tests for two error-cases, happy-path and default-setup. i would like to write some addion

Re: trigger errors

2011-11-14 Thread Ivan Habunek
On 13 November 2011 14:39, Florian Semm wrote: > the configurator and the adapters are good examples for separation of > concerns. :) > > the configurator-test only has tests for two error-cases, happy-path and > default-setup. i would like to write some addional testcases. That would be very wel

Re: trigger errors

2011-11-13 Thread Florian Semm
Am 13.11.2011 10:54, schrieb Ivan Habunek: On 13 November 2011 09:25, Florian Semm wrote: if we expect in a configuration an attribute like "name" or "class" and we got nothing, we should call this trigger_error() function. i haven't seen that the LoggerConfigurator class validate the configur

Re: trigger errors

2011-11-13 Thread Ivan Habunek
On 13 November 2011 09:25, Florian Semm wrote: > if we expect in a configuration an attribute like "name" or "class" and we > got nothing, we should call this trigger_error() function. > > i haven't seen that the LoggerConfigurator class validate the configuration. > The configurator still trigger

Re: trigger errors

2011-11-13 Thread Florian Semm
Am 12.11.2011 20:40, schrieb Christian Grobmeier: On Sat, Nov 12, 2011 at 8:36 PM, Ivan Habunek wrote: On 12 November 2011 20:31, Christian Grobmeier wrote: Good question. In my world logging framework should never break an application. So... what exactly is an error? A console output? stand

Re: trigger errors

2011-11-12 Thread Christian Grobmeier
On Sat, Nov 12, 2011 at 8:36 PM, Ivan Habunek wrote: > On 12 November 2011 20:31, Christian Grobmeier wrote: >> Good question. >> >> In my world logging framework should never break an application. So... >> what exactly is an error? A console output? standard php logfile >> problem? > > What I di

Re: trigger errors

2011-11-12 Thread Ivan Habunek
On 12 November 2011 20:31, Christian Grobmeier wrote: > Good question. > > In my world logging framework should never break an application. So... > what exactly is an error? A console output? standard php logfile > problem? What I did in the new configurator is use trigger_error() with E_USER_WAR

Re: trigger errors

2011-11-12 Thread Ivan Habunek
On 12 November 2011 20:26, Florian Semm wrote: > i have ran some unit-tests with an incomplete xml-config. so how exact > should be an configuration-error? > > my appender config looks like this: > > > > there is no class attribute or layout option. should all this missing > options trigger an er

Re: trigger errors

2011-11-12 Thread Christian Grobmeier
On Sat, Nov 12, 2011 at 8:26 PM, Florian Semm wrote: > i have ran some unit-tests with an incomplete xml-config. so how exact > should be an configuration-error? > > my appender config looks like this: > > > > there is no class attribute or layout option. should all this missing > options trigger

trigger errors

2011-11-12 Thread Florian Semm
i have ran some unit-tests with an incomplete xml-config. so how exact should be an configuration-error? my appender config looks like this: there is no class attribute or layout option. should all this missing options trigger an error? regards florian