Forums : Technical Issue Help
Dear Open Hub Users,
We’re excited to announce that we will be moving the Open Hub Forum to
https://community.blackduck.com/s/black-duck-open-hub.
Beginning immediately, users can head over,
register,
get technical help and discuss issue pertinent to the Open Hub. Registered users can also subscribe to Open Hub announcements here.
On May 1, 2020, we will be freezing https://www.openhub.net/forums and users will not be able to create new discussions. If you have any questions and concerns, please email us at
[email protected]
some CVS commits are not picked up
Hello,
I just noticed ohloh did an incremental pass over the CVS
repo of MARF (http://ohloh.org/projects/3508/commits) today, but
apparently it missed at least one commit. On Monday, December 11
I ('mokhov') had two separate commits done [1] and [2], but ohloh picked
up only [1]... I wonder if possibly some other commits might have been
missed in the past.
[1] Ohloh captured this commit on December 11:
http://marf.cvs.sourceforge.net/marf/marf/src/marf/FeatureExtraction/FeatureExtraction.java?view=log
http://marf.cvs.sourceforge.net/marf/marf/src/marf/Storage/FeatureSet.java?view=log
[2] Ohloh missed this commit on December 11:
http://marf.cvs.sourceforge.net/marf/marf/src/marf/Classification/Distance/MahalanobisDistance.java?view=log
Thanks,
-s
I also noticed a duplicate commit record, where the same commit got two IDs:
3514681 mokhov Sun Oct 15 22:24:41 UTC 2006 Fix the bug #1556320 of FeatureExtractionAggregator dealing with
2643951 mokhov Sun Oct 15 22:24:41 UTC 2006 Fix the bug #1556320 of FeatureExtractionAggregator dealing with
-s
Hello mokhov,
Thanks for help in finding this problem and providing the (very helpful) links. There's a lot going on here.
Over the last couple of weeks we've been burning a lot of energy contending with some database corruption problems. I believe these problems are now solved. I thought we had repaired all of the corrupted projects, but evidently not. I've scheduled an entirely new, fresh download of the MARF code, and I'll make another review our database for similar problems in other projects. This is a pain, but it's a solved problem.
I'm more concerned about the missing
commit on December 11 for MahalanobisDistance.java, because I think our system did all the right things but didn't get the expected answers.
From our logs, it looks like our download engine indeed tried to download the CVS patch for Mon Dec 11 22:56:34 2006 UTC. This is the missing commit you identified. Strangely, it appears that the CVS server returned an empty patch set in response to this request. As a result, we (correctly, I believe) ignored the commit, because we believed it contained no changes.
The next activity on this module occured at Thu Dec 28 06:01:57 2006 UTC, and when our downloader fetched the patch for this CVS revision it received not only the expected new code for that revision, but also received the delayed change for MahalnobisDistance.java that was not returned before. Now disaster has stricken: while we've caught up on code changes, we've misattributed the time and comment of this checkin, and possibly its author.
Was this a one-time error? Today, I could not reproduce the problem manually. It does feel to me from our logs that we attempted to do the correct thing and the CVS server behaved strangely, but I can't guarantee this. When the new download for MARF completes I'll take another look.
I'll keep you posted, and thanks for the help,
Robin
Yeah, thanks for the explanation! It seems to have finished the re-processing of the marf module and so far all looks good, even the duplicate I mentioned earlier is gone. Mustbe there was some race and non-atomic updates somehow creepped in in a network lag or something...
Thanks again!
-s