PCL Developers blog

Michael Dixon

This is my personal page.

Project:Point cloud registration

About me

I’m a research engineer at Willow Garage. I work on the Point Cloud Library.

Recent status updates

Including citations in a blog post
Monday, June 13, 2011

We recently encountered a problem with the way our blogpost extension was handling citations, but I tracked down the problem, and I believe I’ve got it fixed now. It should now be possible to cite an article [DixonCVPR2011] from within a blogpost, as long as the citation target is also defined in your post. If you want to see an example, take a look at the source for this post here.

[DixonCVPR2011]On analyzing video with very small motions, Michael Dixon et al.
Bug fixes
Tuesday, May 31, 2011

Today I fixed a couple of bugs in our blogpost extension, and I’m writing this post as a test.

First, Alex reported that he was unable to include images inside of a .. blogpost::. The problem was solved by adding in a call to process_images when creating the blogbody content.


Second, Gabe found that the :doc: role failed to work when included in a post, and this turned out to be a similar problem. The solution was to add a call to resolve_references.

Finally, Pararth was unable to link to downloadable files, but that appears to be working now, too.

I suspect I’ll need to make a few more fixes along these lines, but I’m afraid that will have to wait until another day. If any of you run into any more problems, please let me know.


Apparently, there’s one more problem I’ll need to solve to get :doc: and :download: working. It looks like those commands are assuming our site’s root is “pointclouds.org/” instead of “pointclouds.org/gsoc/”, so all of the automatically generated links are missing the “gsoc/”. I’ll have to look into that when I have a chance...

Update 2

Fixed. The problem was that a path defined relative to, say, pointclouds.org/gsoc/mdixon/ wouldn’t be valid in a blogpost that was copied over to poinclouds.org/gsoc/index.php. The resolve_references function is designed to handle that, but it didn’t follow the same conventions as the process_images or process_downloads, so I needed to pass in a different “docname” in order to get the right behavior.

How to install Sphinx (and some other stuff)
Friday, May 20, 2011

In my earlier post, I discussed how to write and publish status updates using Sphinx and svn. However, as you might imagine, before you can follow those instructions, you’ll need to install Sphinx (and a couple of additional dependencies for the extensions we use). Assuming you don’t already have them installed, here’s what you need to do:

First, install Sphinx 1.0.7. I have Python 2.6, so this is the one I installed:

~$ easy_install http://pypi.python.org/packages/2.6/S/Sphinx/Sphinx-1.0.7-py2.6.egg

Next, install python-dateutil. (Used by our blog extension to parse the date strings)

~$ easy_install http://labix.org/download/python-dateutil/python-dateutil-1.5.tar.gz

Finally, install dvipng. (Used by the pngmath extension to generate LaTeX formulas)

~$ sudo apt-get install dvipng

After you have these installed, you should be able to “make html” (see earlier instructions) without problems.

P.S. As you can tell, these instructions assume you’re working in Linux. If you’re planning to develop in Mac OS X or Windows, then installing dvipng may be a little more complicated for you. (I haven’t tried to install it on Windows, but this page suggests that “The quickest way is to download and install MikTeX - http://miktex.org/2.9/setup ”, so you might try that.)

See all status updates