| posted in: nerdliness 

Update September 11, 2011 As interesting as figuring out all of the ins and outs of pagination in Octopress was, the real issue that caused my problems was an incomplete update cycle. When you update Octopress you first pull changes from Octopress and then there are two rake commands to run, one to update styles and one to update source templates. I neglected to do the rake commands after a pull and as a result I got plugin changes without the necessary source and styling changes to make them work properly. For a complete look at updating Octopress see the Updating Octopress section of the documentation.

One of the risks when changing how your site is published is breaking links. Links from external sources and links internally from post to post. Toward that end I’ve been making great use of Mint and it’s many Peppers to track visitors and any errors they may be experiencing.

After a recent source code update from Octopress I started seeing some errors referencing URLs that looked like “page/#/”. Last night I discovered pagination on my site was well and truly broken. Whoops.

Pagination is the process of taken all my many posts (1793 counting this one) and spliting them up into pages containing 10 or fewer posts. Since weblogs are published in reverse chronological order, each new posting causes a ripple through all the pages of a site. In my case there are 180 pages of prior posts to re-create. Each one linked to it’s nearest siblings via “older” and “newer” links.

There are many ways to structure a website. In my case I’ve choosen to have my primary page at the root of my domain. Others choose to house their site in a /blog subdirectory. By default Octopress assumes you fall into the later cateory, and want all your postings to live in a /blogs subdirectory. Since I don’t want that I’ve had to tinker with settings in the _config.yml file and also with the Liquid tags that render the primary index page.

By default the primary index page for an Octopress site wants to build an archives link that leads you to /blogs/page/#. It was easy enough to remove the /blogs portion of that link. The same piece of code was also creating pagination links that were incorrect, by embedding an extra /page in each and by adding an extraneous / at the end of the URL. After some debugging and testing on my site this morning, I produced the following code, which renders the three pagination links properly for my site structure.

  {% raw %}
    <nav role="pagination">
        {% if paginator.next_page %}
          <a class="prev" href="{{paginator.next_page}}">&larr; Older</a>
        {% endif %}
        <a href="/archives">Blog Archives</a>
        {% if paginator.previous_page and > 1 %}
          <a class="next" href="{{paginator.previous_page}}">Newer &rarr;</a>
        {% elsif paginator.previous_page %}
          <a class="next" href="/">Newer &rarr;</a>
        {% endif %}
  {% endraw %}

This code also addresses a Liquid error I was seeing where the “Newer” pagination link wasn’t getting rendered at all. The original code was testing paginator.previous_page to see if it was greater than 1 and causing a “comparison of String with 1” error to be thrown. By changing the code to use the comparison works properly.

Author's profile picture

Mark H. Nichols

I am a husband, cellist, code prole, nerd, technologist, and all around good guy living and working in fly-over country. You should follow me on Mastodon.