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.synopsys.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]
Project hasn't updated in over a week
https://www.ohloh.net/p/buddypress-media
Carl,
Rescheduled the failed update which finished OK (10 added commits). The remainder of the enlistments are 10 hours old. Analysis is now at August 22, 2012 on code collected on August 21, 2012 and last commit is at August 21, 2012 - about 22 hours ago.
Please be sure you remind me here if the project stalls again.
Thanks!
Whatever you did made our project LOC jump by over 30,000 lines. We haven't added that much code during the past week.
Was Ohloh just ignoring one of our enlistments or something?
^C^
Ahh. Here we go ...looks like there's a defect in Ohloh.
When I go to Browse Code
at this URL:
see how there are the folders phpunit-portable
, qunit
, and testlib
in the folder?
Those aren't in our SVN. They were removed almost a week ago. You can verify it at this link:
https://code.google.com/p/buddypress-media/source/browse/#svn%2Fbp_media%2Ftrunk%2Funit-test
^C^
Carl,
If I had known that this would come up I would have taken a before
printout of the project page for comparison. Now, it's too late. I didn't do anything particularly unusual but just restarted the job. We get the (Not a valid URL)
complaint off-and-on when the repository hiccups or some internet outage gets in the way. When we come back to it the problem is almost always cleared and the job finishes OK. Pretty standard stuff.
Now, as to the references to old folders on the ohloh code side, we do updates through incremental code dumps and sometimes we can lose track of overall changes in structure, usually with an error message attached and a failed update. If you could tell me which repository had the directories that shouldn't be there, I can do a fresh re-fetch of that which should clear the problem. Just don't want to do a re-fetch of everything without a need. It's also possible that the results from the ohloh code side are cached and I'm not sure how often it is refreshed from our records but I'm willing to do the re-fetch to be sure and also see if it has an immediate effect. A learning experience...
Thanks!
Carl,
Stupid me... I can see that the repository which bears the mysterious extra directories is listed right there on the screen. Apologies. I'll start a re-fetch immediately.
Thanks!
Carl,
I also note that sometimes unusual jumps in LOC count aren't from code written by the project but packages that have been recently added from other projects and aren't being ignored as they maybe should. Sometimes an ignore order will prevent distortion of LOC counts or distractions in the Languages
section. Often users will swear up-and-down that their project doesn't have any JavaScript, for example, and yet they have overlooked that required copy of YUI Library that is being faithfully counted along with everything else.
Continuing the re-fetch.
Thanks.
In this case it's not a package. We keep all our external packages in a /lib folder that isn't indexed by Ohloh.
^C^
Carl,
Cool! One less thing to worry about.
Halfway through the fetch stage with three steps more to go.
Gotta say ...you guys must have a tough job. You take everything that's 'hard' about search and then stack it on top of connecting to hundreds of thousands of unreliable version control systems and updating the whole thing in more-or-less real time.
Hey while you're at it, can you verify that my language stats are updating correctly:
https://www.ohloh.net/accounts/carlroett/positions/total
And also, since I have nearly a million-line delta across 3 different projects with 2.5K commits, that my 'kudo' rank is updating properly too. I'm an 8. Other people on our team have 50K delta and 100 commits and they're ranking at a 7.
^C^
Carl,
Finished with the re-fetch and analysis. No change to the ohloh code side so I suspect that the ohloh code analysis is done periodically and not sure yet how to force a fresh look. Will find out more. LOC for the project is down overall by quite a bit this time though (8/2012 was 147,707 now 127,173 for code etc.) but the graph is still tilted very up in the last month and perhaps more so than before. (was +21% now +33%)
Any other very active enlistments that need a re-look? Where does the most development activity occur? The big branch merges?
Thanks!
We just split our project into two different projects, so other developers can use parts of our project in their projects ...pretty much everything needs refreshing.
Try refreshing Razor as well to make sure everything's working. It runs on the same SVN servers. https://www.ohloh.net/p/wp-razor
^C^
Also, looks like those ghost branches are still in Ohlo's copy of the SVN:
http://code.ohloh.net/project?pid=TwWPiw4gMDA&browser=Default&ipid=301673&cid=-pFJy70y0so&prevcid=&prevDid=
^C^
Carl,
Those are the things we don't always handle as well as we could, the massive re-arrangements. I'm surprised we haven't had more failures... Will do.
The ohloh code issue is something I'll need to take up after 9AM eastern US time when the crew comes in. Another hour or more, probably. I need some additional training, I'll wager.
Thanks!
Awesome. Well, we're at it 24 hours a day, so if you need some debugging info or want us to do something to one of the repos, just let me know.
^C^
Carl,
The re-fetch of everything is done for both BuddyPress Media
and Razor Unit Test Platform
. No particular change to Buddy Press' analysis is seen. The numbers vary very slightly but overall are in line with the last analysis. There is that 33% increase in lines of source code August over July which I don't have any explanation for. Compare to 6% increase July over June and almost the same June over May. I might guess there is some duplication going on to explain it with an extra copy of some code being held somewhere we are analyzing rather than fresh code production. You can experiment at your leisure with toggling on and off the enlistments in history then waiting for the analysis to run. That might reveal which enlistment the increase is coming from. I also wonder about the wiki enlistment: could there be something hiding there? Does it contribute any useful source code to the counts?
Thanks!
It looks like you have some kind of caching issue going on. Browse code
is still showing the ghost repo branches, even when I remove the 'unit-test' branch from the enlisted repositories. Some code analysis screens are showing correct data while others are missing data. 300K lines of code disappeared from my language stats.
Its almost as if Ohloh got 1/2 way through analysing our repos and counting-up lines, then the process died.
^C^
Carl,
I can assure you that the original ohloh side of things is working properly - no failures and with completely fresh fetches for all enlistments. You are correct in your observation that the ohloh code data for the project is not dynamic. It is currently in beta and all the bells and whistles are not yet connected. I will ask that the data for your project be updated. Hopefully that can be done in a day or so.
Sorry about any confusion.
Thanks!
Carl,
"Subject: BuddyPress Media project has been reindexed and published on ohloh code
The BuddyPress Media project should be up to date. Per the statistics it picked up and indexed 707 files, so we certainly picked up a number of changed/moved files."
It looks OK to me but you will know for sure.
Thanks!