> hackystat-mds (because it's also cool) (used to be jpl) probably should be "hackystat-sqi", because the hackyApp_Sqi module supports other harvest stuff. At least that is what I think I did.
>> 8. emma: troublesome, see separate email. > OK, keep working. As anyone seen the Emma println that I said to check on the tomcat console when running from cruisecontrol? > > 10. locc: > > Locc is working. There was a bug in locc sesnor, but Mike fixed it Friday, > > not sure the fix is on hackydev or not. > Excellent. FYI: Locc doesn't entirely work with Java5. There are some java 5 syntax that it doesn't support. thanks, aaron ----- Original Message ----- From: Philip Johnson <[EMAIL PROTECTED]> Date: Monday, November 21, 2005 1:30 pm Subject: [HACKYSTAT-DEV-L] Summay of work on hackydev > --On Saturday, November 19, 2005 2:49 PM -1000 "(Cedric) Qin ZHANG" > <[EMAIL PROTECTED]> > wrote: > > > This is a summary of work on hackydev. > > > > 1. CVS sensor is taken down. SVN sensor is up. All previous svn > commit data have been > > sent to hackystat server. I also configured the sensor to run > nightly.> - SVN sensor supports file rename tracking. There > would be no more bogus churn > > unless you are not using svn rename facility. > > - SVN sensor runs much faster than CVS sensor, because it does > not have to scan all > > the files in the repository. > > - SVN sensor also sends commits information about binary files > to hackystat. For > > binary file, total lines, lines added/deleted are all zero. > > - SVN sensor code is completely rewritten, Todd's contribution > is not used . > > Excellent work! > > > 2. V7 Cruisecontrol supports multiple hackystat configuration, > though currently only > > hackystat-all is built. A new configuration can be added in less > than 5 min. > > > > Philip, could you tell us which configuration you want to > support? I guess we no longer > > need to build hackystat-Jpl and cocomo. > > Let's have: > hackystat-all > hackystat-uh (no longer including cgqm) > hackystat-cgqm (just what is needed for cgqm) > hackystat-standard (minimal config, no zorro, etc.) > hackystat-cocomo (because it's cool) > hackystat-mds (because it's also cool) (used to be jpl) > > > Following are hackydev nightly build individual task status: > > > > 1. The distAll (in cruisecontrol.build.xml) task does not zip > files correctly. Thus the > > distribution files for each night's integration build are not > really working versions. > > I'll take a look. In 7.0, the naming scheme for targets changes a bit. > > Use "all.<target>" to indicate a target that calls local targets > (such as all.junit, > which calls hackyCore_Kernel.junit, etc.) > > Use hackyCore_Build.<target> to indicate a target that does not > call local targets. > > Do not use <target>All to name any targets. If you see those, you > should rename or > delete. > > > 2. Svn change log, no support right now. > > > > Only "svnant" provides ant interface to svn, but the problem is > that it does not > > support "svn log", which is required to get the log (i.e. which > files are committed for > > the previous day) information. > > > > There are 2 options left: > > > > (a) Call svn native client from ant and ask it to generate a > log file in xml formate. > > We can write a xslt file. With this approach, we can get commit > information from "the > > last revision of the day before yesterday" to "the last revision > of yesterday". > > > > (b) Write a "svn log" ant task ourselves using the same library > svn sensor uses > > (estimated work load: 3 days). This way, we can get commit > information from "the first > > revision of yesterday" to "the last revision of yesterday". > > Since we have the hackystat-svn-l mailing list, this report not as > important as it was. > Let's just put it aside until a very simple solution comes along. > > > > 3. checkstyle: fully working. > > excellent. > > > > > 4. hackyPerf, not there yet. > > OK, keep working on it. > > > > > 5. java2html: fully working > > > > 6. javadoc: fully working > > Excellent. > > > > > 7. jblanket: incompatible with java5 > > OK, let's put this aside until if/when jblanket supports Java5. > > > > 8. emma: troublesome, see separate email. > > OK, keep working. > > > > > 9. junit: fully working except > > > > The junitReport is written in such a way that it only generates > report when there is > > junit failure. So, if junit is successful, we see nothing on > hackydev. > That's OK for now. > > > > > 10. locc: > > > > Locc is working. There was a bug in locc sesnor, but Mike fixed > it Friday, not sure the > > fix is on hackydev or not. > > Excellent. > > > > > 11. docbook: > > > > The buld script does not work. > > I will work on this. > > > > > Cheers, > > > > Cedric > > Thanks, Cedric! > > Cheers, > Philip >
