Sitemap logic needs to conform to specs.

[chrissound]

A quick google search for "haskell persistent insert" seems to return the docs for v0.3.1.3 (latest is v2.7).

I think if we set a higher priority for the latest version in the sitemap.xml - Google will pick this up and have the latest version higher up in the search results.

[hvr]

Here's what the sitemap says currently:

  <url>
    <loc>https://hackage.haskell.org/package/persistent/docs</loc>
    <lastmod>2017-04-17</lastmod>
    <changefreq>daily</changefreq>
    <priority>1.0</priority>
  </url>

  <url>
    <loc>https://hackage.haskell.org/package/persistent-0.3.1.3/docs</loc>
    <lastmod>2017-04-17</lastmod>
    <changefreq>monthly</changefreq>
    <priority>0.25</priority>
  </url>

So it does already use a higher priority for the unversioned URL which always points to the last version...

[chrissound]

Hmm. The v2.7 entry is at a priority of 0.25 though.

(Please excuse the strange formatting).

><priority
>0.25</priority
></url
><url
><loc
>https://hackage.haskell.org/package/persistent-2.7.0/docs</loc

[hvr]

True, but if google already now doesn't appear to prefer persistent/docs over persistent-0.3.1.3/docs, why would it make a difference if the priority for persistent-2.7.0/docs would be inbetweeen? (I'm not saying we shouldn't try it, I'm just trying to understand why you think it would make a difference).

[chrissound]

I'm honestly not too sure, I'm just taking a guess here.

[gbaz]

I finally got us set up on the search management console and discovered that our sitemap is rejected by google :-/

"Your Sitemap contains too many URLs. Please create multiple Sitemaps with up to 50000 URLs each and submit all Sitemaps."

So we'll need to rethink some logic here.

[gbaz]

actually its not rejected -- it just only accepts that first chunk of urls, which is the majority...

[gbaz]

we can do multiple sitemaps as per: https://support.google.com/webmasters/answer/75712

Also sitemaps should have a link for every url, not just subdirectories.