Final part!
Also, it is really, really, really strange to find TestProjectManager in the
activity.dailyanalysis package.
Yes, I agree. Actually, the root problem is that:
- The "Project" definition code is in the "Core" subsystem.
- The Project definition test code requires you to define a Project, which requires a
Workspace, which requires sensor data, which requires at least one SDT declaration, which
requires a hackySdt_* module, which creates a circular dependency (Core <-> Sdt).
The way I resolved this for 7.0 was to not test the Project code in the Core subsystem,
but to move those tests into a module in the SDT subsystem, where there would be sensor
data available with which to test things. That breaks the circularity, but creates the
problem of a test case "appearing out of nowhere" in an SDT module.
In 7.1, I will fix this by creating data for the EvolSdt and using that for "Core"
subsystem tests:
<http://hackydev.ics.hawaii.edu:8080/browse/HACK-431>
Action Items:
1) Add an Ant dependency on hackySdt_Activity in the
hackySdt_FileMetric/build.xml
Done.
2) Change the javac task to use fork in the hackyCore_Build/build.xml
Fixed by setting includeAntRuntime=false
3) Change the class name in
org.hackystat.core.common.dailyanalysis/dailyanlaysis.basic.xml file.
Fixed by deleting dailyanalysis.basic.xml
4) add MostActiveFile, MostActiveFileMetric, MostActiveFileWorkspace to a
dailyanalysis.<name>.xml file.
Done.
5) Add Test Cases in the dailyanalysis packages, that actually tests the Web
Interface.
Scheduled for 7.1:
<http://hackydev.ics.hawaii.edu:8080/browse/HACK-432>
ps. I was thinking that it would be cool if my sensor data would indicate that
I was
debugging a problem. I have little to no active time, but I sent a bunch of
Build,
Unit Test, and Activity data (a lot of code reading and browsing). I spent a
couple of
hours tonight figuring all this out and wanted to know how much time I spent in
the
past figuring this kind of stuff out. Any ideas?? Maybe Hongbing's micro
development
streams can discover "debugging".
Yes, that would be an excellent thing to try to do with the upcoming "DevEvent" SDT and
software development stream analyses.
Cheers,
Philip
