It looks good to me. I've checked in your changes. > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:nant-developers- > [EMAIL PROTECTED]] On Behalf Of R. Owen Rogers > Sent: Tuesday, October 01, 2002 11:09 AM > To: [EMAIL PROTECTED] > Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]; > [EMAIL PROTECTED]; [EMAIL PROTECTED] > Subject: [nant-dev] XmlLogging mods, Part 2 > > We are in the process of finalizing the development of a functional port > of > the continuous integration tool CruiseControl > (http://cruisecontrol.sourceforge.net) to the .NET platform. In doing so, > we have made a number of enhancements to the XmlLogging capabilities of > NAnt > that were originally implemented and posted by Bill Caputo. I have > confirmed all our modifications with Bill, and hopefully these changes > won't > affect anyone else. > > We have made 3 basic changes: > 1) We have changed the output of the XmlLogger from > <foobar><exec/></foobar> > to <target name="foobar"><task name ="exec"></target>. This helps > identify > which elements are targets and which are tasks. More importantly, it > makes > browsing the Xml output using XPath much more explicit. The generated > output is also now very similar to what is produced by Ant. > > 2) We needed to be able to identify the type of messages generated by > particular tasks. Specifically, we needed to identify which messages > generated by the compile tasks were compiler warnings and errors and which > were just informational messages. In Ant, this is solved by using a > priority attribute attached to different messages. To accomplish this in > NAnt we added a WriteMessage method to Log. Unfortunately we couldn't > address the issue by overloading the Write or WriteLine methods, so we had > to add a new method. The WriteMessage method adds a type attribute to the > outputted message element. Currently it does not modify the output for > non-xml loggers. > > 3) We modified the NUnit task to redirect its Xml output to the console if > required. The NUnit task provided a usefile attribute, which was actually > not used (if it was activated the output would just disappear -- ie. it > would > not be logged to file). We modified its implementation so that it would > redirect its output via the logger to the console. To this effect, we > created a LogWriter class which subclasses StringWriter. It uses a string > buffer to cache the output and then redirects it to the Log when the > stream > is closed. This class enables us to swap implementations that expect a > file stream for logging output to file. The nunit2 task, when it is > ready, > should work similarly. > > Hopefully these modifications are all acceptable. As a point of > reference, > I've attached the original email from Bill to the list below. > Best regards, > Owen.
------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Nant-developers mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/nant-developers
