Forum Posts

Ohloh uses an open source tool to detect and count lines of code. https://github.com/blackducksw/ohcount When I run this tool against the code in the repository it's not detecting the files as ... [More] fortran. Being open source if you'd like to fork ohcount and enhance it to find the flavor of fortran in sdsu-asklib, we can then incorporate those changes into Ohloh. I've opened a ticket on our side to look into fixing this at some point in the future. [Less]

Geoffrey: You are now the sole manager of the jboss-drolls project, OhlohBee is no longer one.

Hello Ohloh automatically trusts server certs, so that's not the issue in this case. When running the following command: svn log --xml --stop-on-copy -r 1:HEAD ... [More] '[[https://svn.dotclear.net](https://svn.dotclear.net)/lab/@HEAD'](https://svn.dotclear.net/lab/@HEAD') --no-auth-cache The entire log is not sent. Instead it's truncated and the server send tacks on at the end an erroneous line: svn: REPORT of '/lab/!svn/bc/3027': 200 OK (https://svn.dotclear.net) This is a common SVN problem and searching Google for: "200 OK" svn ... yields many discussions about this problem. In some cases it's due to a post-commit hook on the server failing. In other cases it's due to a bungled CVS -> SCN conversion or a file corruption problem. Until the command near the top of this comment runs correctly, Ohloh won't be able to update the project. [Less]

Bastie: You are correct. There is a bug in our detection. Ohloh uses a tool called 'ohcount' (open sourced here https://github.com/blackducksw/ohcount) This tool is reporting several .cs files as ... [More] java. I have opened a bug on our side about this so we can investigate why this is happening and fix it. [Less]

Something we developers here at Ohloh really want to do is improve the speed/performance of Ohloh's back end crawling/analyzation code. We want it to be able to handle large projects with a long ... [More] commit history without needing personal attention. Sorry for the problems. [Less]

Since my last update, we've fixed a few critical Ohloh bugs. We've also fixed one of the two major problems I mentioned, incorrectly counting source code for GIT repositories. Over the past 2-3 ... [More] weeks or so we have been examining Ohloh's back end and how we can modify it to scale better. We are currently not automatically adding projects from Github or Code.Google and we'd like to do this but we need make sure the back end can handle it first. So no progress has been made on adding darcs support, sorry. [Less]

So, Robin knows a lot more about the Ohloh back end than I do. Also, I wasn't aware of how many revisions are involved. I also wasn't aware of the length of time it took per revision. So Robin's ... [More] conclusion stands, it won't be possible to import this into Ohloh at the current time. [Less]

We've identified code in the back end that was slowing all our job processes way way way down. We're easily running 10x as many jobs as we were before. I've re-scheduled the Gentoo CVS job and hope ... [More] that this time it runs faster thanks to our back end improvements. I'll keep an eye on it over the next day or two and let you know how it goes :) Also, it looks like the bazaar repository is broken due to a transient Ohloh bug. I'm going to work on fixing that as well. [Less]

lars: We are seeing a weird error whenever we try and count the code: abort: /home/deployer/audit.c not under root We're not sure why this is. It requires some investigation.

The project wasn't updating due to a broken enlistment. This caused the badge to remain visible. Now that the project has updated, the badge is now gone.