Unexpected urls in my Jekyll website sitemap - SEO

davioooh
August 14, 2017
230 views
3 votes
4 Answers

I’m creating my personal blog (davioooh.com) using Jekyll and hosting it on GitHub pages.

But I recently discovered a strange thing…

When pushed/deployed to my repository, the blog sitemap.xml includes two extra urls:

<url>
  <loc>
    http://davioooh.com/assets/javascript/anchor-js/docs/grunticon/preview.html
  </loc>
  <lastmod>2017-08-11T16:40:49+00:00</lastmod>
</url>
<url>
  <loc>
    http://davioooh.com/assets/javascript/anchor-js/docs/index.html
  </loc>
  <lastmod>2017-08-11T16:40:49+00:00</lastmod>
</url>

I can’t understand why… I’m not using any anchor-js gem or plug-in…

The only gems declared in my _config.yml and Gemfile are:

gems:
  - jekyll-feed
  - jekyll-seo-tag
  - jekyll-paginate
  - jekyll-sitemap

So why I’m getthing that urls?

NOTE: the sitemap is correct when I test the site locally on my dev machine.

NOTE: I’m using a custom theme (regular theme, not gem-based). No theme declared in my _config.yml. Here you find my blog repository: https://github.com/davioooh/davioooh.github.io

UPDATE Tried to replace jekyll gem with github-pages in my Gemfile as suggested by marcanuy. It works locally, but after a new push sitemap still includes the extra urls…

UPDATE Found GitHub pages reference listing default params for Jekyll websites. Default theme is, as expected, jekyll-theme-primer. How can I override this setting if I’m not using gem-based theme?

Answers

Chosen as BEST ANSWER
- davioooh
- August 28, 2017 at 9:46 pm
- 0 votes
0
It seems that updating to github-pages v157 the sitemap is generated correctly.

The new version of gh-pages supports the last version of jekyll-theme-primer (v0.5.2).

This solves my issue.

(Edit)

- ashmaroli
- August 14, 2017 at 11:57 am
- 0 votes
0
It comes from your theme: jekyll-theme-primer

Primer is what powers GitHub’s frontend. If you’re not using this theme explicitly, it probably comes from GitHub Pages’ own usage of the Primer framework

Login or Signup to reply.

- marcanuy
- August 14, 2017 at 1:52 pm
- 0 votes
0
Why it works fine locally?

Because you are not using the same dependencies Github Pages uses. To replicate the environment used by Github, you need to use the gem: github-pages.

Why sitemap.xml includes two extra urls?

As @ashmaroli said, these files comes from jekyll-theme-primer which is used by Github Pages, if you are using your own theme files, then just make sure there are no theme: key in your _config.yml and run bundle update. Using the right github-pages gem will fix it and those files won’t be included in your final site.

Login or Signup to reply.

- FrankN
- April 13, 2018 at 9:27 am
- 0 votes
0
1. Apparently, the jekyll-theme-primer (which usually comes with group: :jekyll_plugins in your Gemfile) is to blame. I.e. theme: minima avoids the creation of anchor-js (but you probably want to run your own theme, and not having minima hang around…)

2. A verified simple fix (giving a warning upon build) is to write
```
theme: false
```
in your _config.yml

3. The issue is a known issue and has been fixed 3 days before me writing this. Might day some days (or weeks) until it gets shipped though. If you read this some time from now, probably all you need to is update your gems.
Login or Signup to reply.