[gatsby-source-wordpress] fetchReferencedMediaItemsAndCreateNodes resolves too early

[rburgst]

Description

I have an ACF that contains HTML content which contains multiple audio element which is referencing a wordpress hosted audio file. The sum of all referenced media items exceeds the configured schema.perPage configured page size.
In this case it seems that the promise for fetchReferencedMediaItemsAndCreateNodes resolves with only parts of the received data (IMHO the last page) rather than the combined list of media Items.

This leads to the case where you have only a few of the media items urls replaced, while the others still point to the wordpress server.

Steps to reproduce

1. generate the starter project gatsby new my-wordpress-gatsby-site https://github.com/gatsbyjs/gatsby-starter-wordpress-blog
2. configure the following options

    schema: {
         perPage: 2,
       },
   

3. Create an empty wordpress instance, install, wp-graphql and update the automatically created post to contain 3 audio tags (mp3)
4. run

   yarn clean
   yarn develop
   

5. now go to http://localhost:8000/___graphql and execute the following query

    query MyQuery {
     allWpPost {
      nodes {
        slug
        content
        }
      }
    }
   

6. observe the output HTML

     {
      "data": {
         "allWpPost": {
           "nodes": [
             {
               "slug": "hello-world",
               "content": "\n<p>Welcome to WordPress. This is your first post. Edit or delete it, then start writing!</p>\n\n\n\n<figure class=\"wp-block-audio\">
                       <audio controls src=\"http://dockerhost:8000/wp-content/uploads/2021/05/file1.mp3\"></audio></figure>\n\n\n\n<figure class=\"wp-block-audio\">
                      <audio controls src=\"http://dockerhost:8000/wp-content/uploads/2021/05/file2.mp3\"></audio></figure>\n\n\n\n<figure class=\"wp-block-audio\">
                       <audio controls src=\"http://dockerhost:8000/wp-content/uploads/2021/05/file3.mp3\"></audio></figure>\n"
             }
           ]
         }
       },
       "extensions": {} 
    }
   

Expected result

the 3 audio files should be rewritten to /static... urls

Actual result

The urls are not rewritten at all.

Environment

  System:
    OS: macOS 11.4
    CPU: (16) x64 Intel(R) Core(TM) i9-9880H CPU @ 2.30GHz
    Shell: 5.8 - /bin/zsh
  Binaries:
    Node: 14.16.0 - /var/folders/dm/59s6_w2s3jl78c__p8j_cn1w0000gn/T/yarn--1622281277197-0.1413190554991537/node
    Yarn: 1.22.10 - /var/folders/dm/59s6_w2s3jl78c__p8j_cn1w0000gn/T/yarn--1622281277197-0.1413190554991537/yarn
    npm: 7.7.6 - ~/.nvm/versions/node/v14.16.0/bin/npm
  Languages:
    Python: 2.7.17 - /usr/local/bin/python
  Browsers:
    Chrome: 90.0.4430.212
    Firefox: 88.0.1
    Safari: 14.1.1
  npmPackages:
    gatsby: ^3.3.0 => 3.6.1 
    gatsby-image: ^3.3.0 => 3.6.0 
    gatsby-plugin-manifest: ^3.3.0 => 3.6.0 
    gatsby-plugin-offline: ^4.3.0 => 4.6.0 
    gatsby-plugin-react-helmet: ^4.3.0 => 4.6.0 
    gatsby-plugin-sharp: ^3.3.0 => 3.6.0 
    gatsby-source-filesystem: ^3.3.0 => 3.6.0 
    gatsby-source-wordpress: ^5.4.1 => 5.6.0 
    gatsby-transformer-sharp: ^3.3.0 => 3.6.0 

[rburgst]

Note that for repro you should always clear the cache (yarn clean) before starting yarn develop

[rburgst]

created a sample repo (although there is not really anything to it other than setting schema.perPage to 2

https://github.com/rburgst/gatsby-wordpress-file-rewriting-bug

[rburgst]

This is what the default wordpress post with 2 audio elements looks like:

[rburgst]

Note that the rewriting works if I only add a single audio clip. It starts to fail with 2 audio clips.

[rburgst]

I am fairly certain the problem is in fetchMediaItemsBySourceUrl it seems to me that the whole promise is being resolved when the first page is completed. Therefore, the await in replaceFileLinks is only receiving the data from the first page.
The problem is that audio tags in wordpress append a ?_=1 to the URL,

<audio
  class="wp-audio-shortcode"
  id="audio-2519-7"
  preload="none"
  style="width: 100%"
  controls="controls"
>
  <source
    type="audio/mpeg"
    src="https://wordpress.host/wp-content/uploads/2018/05/file1.mp3?_=7"
  />
  <a href="https://wordpress.host/wp-content/uploads/2018/05/file1.mp3">
    https://wordpress.host/wp-content/uploads/2018/05/file1.mp3
  </a>
</audio>

therefore, the graphql query that loads the media info data will produce empty results for https://wordpress.host/wp-content/uploads/2018/05/file1.mp3?_=7 but it will actually resolve the https://wordpress.host/wp-content/uploads/2018/05/file1.mp3 from the <a href> tag.

Thus, the first graphql query page data will contain many null fields and therefore, you need to wait until all pages are done and resolve the concatenated result.

[github-actions]

Hiya!

This issue has gone quiet. Spooky quiet. 👻

We get a lot of issues, so we currently close issues after 60 days of inactivity. It’s been at least 20 days since the last update here.
If we missed this issue or if you want to keep it open, please reply here.
As a friendly reminder: the best way to see this issue, or any other, fixed is to open a Pull Request. Check out gatsby.dev/contribute for more information about opening PRs, triaging issues, and contributing!

Thanks for being a part of the Gatsby community! 💪💜

[rburgst]

I would really appreciate timely feedback to this important bugfix

[github-actions]

Hiya!

This issue has gone quiet. Spooky quiet. 👻

We get a lot of issues, so we currently close issues after 60 days of inactivity. It’s been at least 20 days since the last update here.
If we missed this issue or if you want to keep it open, please reply here.
As a friendly reminder: the best way to see this issue, or any other, fixed is to open a Pull Request. Check out gatsby.dev/contribute for more information about opening PRs, triaging issues, and contributing!

Thanks for being a part of the Gatsby community! 💪💜

[rburgst]

any news here? this is getting quite frustrating having to wait for months for a bugfix to be reviewed.