Ohloh statistics are broken
| Project: | JNode Website |
| Component: | Miscellaneous |
| Category: | bug report |
| Priority: | normal |
| Assigned: | lsantha |
| Status: | closed |
Jump to:
Description
(This is not a problem with our website, but the tracker UI insists that I give a 'project'.)
The Ohloh widget on the website is currently claiming that JNode has 0 lines of code, 0 developers and is worth $0. I looked at the ohloh page for JNode and I cannot see what the cause is.
It might be related to this event.
There are indications that Ohloh doesn't cope very well with large-scale changes ... like the classlib reorganization, so that might be the cause.
Another possibility is that the big classlib binary files in the jnode tree are causing problems.
Perhaps this would be fixed by deleting the "enlistments" and recreating them? If so, it would be a good idea to avoid enlisting the classlib binaries ... if that can be done.
- Login to post comments
#1
I'm working on theohloh issues.
Indeed, they started with classlib6. When I enlisted the classlib6 directory at ohloh then due to how ohloh appears to be working all the commits done to trunk appreared twice and therefore the number of commits for each commiter was doubled in the reports. The explanation of this might be how I created the claslib6 directory: I made a copy of trunk and removed everything that doesn't belong to classlib6 and I also removed the claslib6 suff from trunk. The consequence of this is that classlib6 inherited the complete history of trunk up to the mentioned copy operation. However this is the case also when we create a branch and so if I would enlist any branch of jnode at oholoh the problem would reappear. I expected ohloh to be able to detect this and creaty a valid statistics of the commits. Unfortunately this is not the case.
Then I started to fix the problem by removing the two enlistements and I added the enlistement again for trunk. Ohloh was reacting slowly. The enlistements disapread instantly but for at least for one day the invalid commit history was still there. Then it was removed and those counters reset to 0. Now I'm waiting for the reenlisted trunk to get enqueued by ohloh for processing. If it keeps pending for long I will have to contact them and report the problem.
For fixing the history problem of classlib6 a solution I have in mind is to create a new empty directory for it and import the contents of casslib6 into it. So the history of classlib6 will begin in that moment.
I will disable the ohloh banner on this forum until it gets fixed.
#2
Ohloh is back and shows correct data now.
#3
Looks right to me. Fixed.
#5
Closed manually.