Forums : Posts

Oh - I wasn't aware. Thanks very much, and best of luck in getting things resolved! :)

Hi Lennart, Yes, I agree. We can continue to throw more and more hardware at the problem on our end, but ultimately that won't scale for us, and I think the answer lies in becoming truly ... [More] decentralized. Sometime in the coming months I hope we'll have a chance to rethink how we import source control history and get away from dependence on our central database. Robin [Less]

I would also like to see support for the bzr/baazar vcs system.

I was thinking about translating ohloh.net. At least in France, lot of people barely understand english or are just willing to stick to french, translating ohloh.net would help them. There are also ... [More] funny things to do with dates, currencies and numbers. As for implementing i18n files handling in ohloh, I am already too busy at work. I haven't seriously looked at MediaWiki code for the last 9 months :,( [Less]

Languages are handled by plugins that detect files of a particular language and then parse them and report the number of lines of whitespace, comments and code. If you're interested in getting a new ... [More] language added, check out: http://labs.ohloh.net/ohcount cheers stuart [Less]

Gnus and XEmacs are both stuck in Step 1 of 3: Downloading source code history (Failed) and has been since the recent /. post pointed me here.

Yeah, I can't deny that we should be smarter about scheduling updates, and that some projects should be updated every day and some only once a week (or never). However, the problem we're struggling ... [More] with right now is that there have been a lot of changes to our line counter lately, so we decided to recount a significant amount of the code on Ohloh from scratch. Now we're stuck with a huge backlog. It's a little compounded by the fact that we just got slashdotted, which dumped about a week's worth of new projects on our doorstep in less than a day, so we'll be backed up for a while :-) Robin [Less]

Hi Iulian, The main problem is that Ohloh simply doesn't have a Scala parser. I'm not at all familiar with Scala, but if the syntax is similar to other languages it might be pretty simple to adapt ... [More] Ohloh's parser to read Scala. The parser is open source and can be found at labs.ohloh.net if you want to take a whack at it. If we're looking at the wrong repository, please by all means edit the list of repositories for the project. Ohloh is driven by user edits: if you find mistakes, feel free to fix them. It also looks like we haven't had a successful update of the (wrong) repository in several months. Does this old repository still exist? Robin [Less]

Hi achow, Ohloh does correctly recognize *.jsp as JavaServer Pages (although we don't recognize *.jspf). I did a lot of digging and I think I've found the cause of this. It turns out that the cause ... [More] is rather suspect, and perhaps we need to change some things in our analyses. The cause seems to be that in November of 2006 you added all the code for the TinyMCE editor to your project. This single commit contained about 48K lines of new JavaScript code, which is just barely greater than the total amount of new Java you have added during your time on the project. The commit where this happened is here. Here's where this gets a bit confusing. When we look at your analysis on our web page, both in lines changed and in number of commits, you appear to work mostly in Java. Why did our analysis pick JavaScript? Answer: because when determining your primary language, our analysis engine is looking only at lines added, and does not consider lines removed, and also does not consider number of commits. The lines changed that you see on your report is the absolute sum of both lines added and lines removed. When you look only at lines added, JavaScript is higher. Unfortunately, the individual lines added and lines removed metrics are not visible on your report page, only the total. I kind of think this is all needlessly complicated, and I can't remember why we decided to use only lines added when picking your primary language. Certainly we should look at some kind of net lines, and certainly at the number of commits. By both of those measures, you are a Java developer. I'll put it in our bug list. Thanks for pointing this out, Robin [Less]

Hi Stephen, Ohcount, our language detector, is open source. Here's the home page. We welcome test-driven patches ;-).