Problem confirmed, but I don't have data with me. I will check tomorrow.
Possible cause:
(1) Coverage data for subpackages not sent to server (possible but highly unlikely).
(2) Problem with mapping: class name => JavaClassWorkspaceMapper ==> file name (this can be tricky).
Thanks.
Cedric
Again,
Aaron Kagawa wrote:
Hi all,
I am experiencing a problem with the coverage calculation via workspaces. I've talked to Hongbing and Cedric about this before, but I still don't understand the problem and how it can be fixed. Here is how you can duplicate my problem:
1) go to the extras page on the Hackystat Public website
2) In the Telemetry Direct Streams Display command configure the selectors to have the following attributes:
- Report Type = table
- Project = Hackysta5-UH
- Interval = Day, September 3, 2004 - September 7, 2004
- Telemetry Streams:
- streams coverage = { JavaCoverage("NumOfCoveredMethods", "**\hackyStdExt\src\org\hackystat\stdext\activity\**") };
3) execute the command. you should see a table with 23 and 12 in the celss.
4) now, try changing the Telemetry Streams textbox to contain:
- streams coverage = { JavaCoverage("NumOfCoveredMethods", "**\hackyStdExt\src\org\hackystat\stdext\activity\analysis\**") };
5) execute the command again. you should see a table with all zeros. why is that?
6) maybe that is a fluke. try
- streams coverage = { JavaCoverage("NumOfCoveredMethods", "**\hackyStdExt\src\org\hackystat\stdext\activity\dailyproject\**") };
* this should return all 12s
- streams coverage = { JavaCoverage("NumOfCoveredMethods", "**\hackyStdExt\src\org\hackystat\stdext\activity\sdt\**") };
* returns all zeros (which is wrong!)
So, I checked the coverage.xml data for September 7, 2004 and
- it appears that the class paths associated with those workspaces have coverage data for them.
- I think that Cedric and Hongbing mentioned that coverage uses FileMetrics to map the class path to the workspace. So, I checked the September 7, 2004 filemetric.xml file and it did contain data for all those workspaces. Everything appears ok.
I'm a little confused on what the problem is. Please let me know if you _cannot_ reproduce this error, as I'm able to reproduce it on three separate machines. If this type of error is expected, definitely please clue me in on what is going on.
One a side note: I am also experiencing some differences between coverage reported in Hackystat and coverage reported from the Jblanket report. For example,
- streams coverage = { JavaCoverage("NumOfCoveredMethods", "**\hackyStdExt\src\org\hackystat\stdext\bufftrans\dailyanalysis\**") }; on September 14, 2004 reports 8 uncovered methods. However the jblanket report for the daily build that occurred on the 14th says there are 10 uncovered methods. Again, I looked in the coverage.xml file for the 14th and I found out that bufftrans uses an internal class that has 2 uncovered methods. Could that be causing the discrepancy? But, I don't think this problem i explain above.
thanks, aaron
