Greetings, all,

Just a quick note to let you know that I committed a bunch of changes 
reflecting some progress on HACK-711, HACK-712, and HACK-713.  These were not 
formally scheduled for 7.4 but they're significant bugs so I've decided to make 
them my priority for right now.   I'm not through yet, but here's a progress 
report:

[1] Aaron reported problems getting StackyHack data to show up in 
DailyProjectDetails. To some extent, this was a result of our old friend 
"runTime" vs. "runtime".  It turns out that SCLC/xmldata sends it as "runtime", 
while the FileMetric expects runTime.   I made a temporary fix by altering the 
SensorDataPropertyMap getProperty method to check for both in the case that 
someone was requesting the runtime (or runTime) attribute, but this is clearly 
a special case of the more general problem.

Upon reflection, I suspect that the right long term solution is to make the 
keys in the SensorDataPropertyMap case-insensitive.  This is simple to 
implement and now that we don't have a data field, the changes would be 
restricted to that single class.  It could potentially have a performance 
impact, but I don't think so.  In any case, we have to weigh that against the 
fact that it took me a day of debugging to figure out this problem, and it's 
bound to re-occur with other property keys as well, with similarly unintuitive 
results. 

Does anyone see any problems with this?  If not, I will schedule it for the 7.5 
release. 

[2] There were also some silly problems with the actual StackyHack sensor 
definitions. For example, the svn sensor had an attribute that prefixed the 
files with c:\csdl-svn\, but the actual prefix should have been c:\svn-csdl\.  
This didn't affect sending the sensor data, but did affect the association of 
this data with the StackyHack project.  

[3] A more serious documentation issue with the StackyHack SVN data is that it 
will always show up in the 14-March-2006 Issue data regardless of when the user 
invokes the sensor. This isn't a bug, it's a feature, but it needs to be made 
clearer in the StackyHack documentation.

[4] Finally, the StackyHack system hard-wires the workspaceroot prefixes for 
both the Jira data and the SVN data to c:\svn-csdl.  This is bad, it means any 
other user will have to physically modify the files, or else the Jira and SVN 
data won't show up in their own project.  It's simple to fix (just have an Ant 
property), but I didn't realize the issue until I was fooling around with this 
stuff. 

[5] Julie reported problems with missing column headers, and more generally, 
incorrect FileMetric totals for hackyCore_Installer.  I'm working on this issue 
as well. I've been able to reproduce the missing column header problem, and 
noticed that it occurred for a day in which there was both SCLC and LOCC 
FileMetric data.  So, I think that is an interesting avenue to explore.  Note 
sure yet whether it is also a cause of the hackyCore_Installer data problem.

None of this is particularly rocket science, but it does take some time to 
track down and then adjust all of the documentation appropriately.  At this 
point, I'm anticipating that I will call it a stable release after fixing these 
issues and leave the chapters for 7.5. 

Let me know if you have any comments or concerns.  We'll probably need at least 
one additional release candidate later this week on the public server. 

Cheers,
Philip 

Reply via email to