I have removed the offending line in trunk. Regards, Ivan
2011/10/3 Ivan Habunek <[email protected]>: > On 3.10.2011. 7:54, Christian Grobmeier wrote: >> >> The only thing to think about is backwards compatibility. Two options: >> >> 1) don't care on bc and remove the line >> 2) add an optional param to createObject, like: createObject($class, >> [$usingNamespace]); > > 3) Stop using createObject because it only obfuscates code? :) > > I'm not a fan of that method. Why is: > $obj = ReflectionUtils::createObject($class); > better than: > $obj = new $class(); > ? > And it can silently do nothing if $class is empty (as twitter users would > say: #fail). :) > >> I would prefer #1, b/c would introduce a param for a php version we >> currently don't support. >> >> Ivan, what do you think? > > I agree that we can do #1 right away. All tests pass without problem. > > Anyway, basename() is meant to be used on file paths, not on class names. > The only reason it works is because "\" is the delimiter for both namespaces > and the windows file system. > > Florian, just a heads-up. I'm in the middle of rewriting the whole > configurator part of log4php. In other words, there will be substantial > changes, and soon. If you're interested in a preview have a look here: > https://svn.apache.org/repos/asf/logging/log4php/branches/experimental/config-adapters > > Regards, > Ivan >
