#4469: Pass Drupal Http Client to Harvester

[stefan-korn]

Fixes #4469

Describe your changes

Just passing the Drupal HTTP Client to the Harvester. If you have suggestion, like do it with dependency injection in the harvester service, let me know.

QA Steps

• Add manual QA steps in checklist format for a reviewer to perform. Be as specific as possible, provide examples if appropriate.

Checklist before requesting review

If any of these are left unchecked, please provide an explanation

• I have updated or added tests to cover my code
• I have updated or added documentation

[stefan-korn]

okay, codeclimate says it all ;-)

\Drupal calls should be avoided in classes, use dependency injection instead

Will take a look into this.

[dafeder]

@stefan-korn yes we would need to add http client as a constructor for the service. Codeclimate will surely complain that it's too many arguments but that one we can ignore; I think the solution there is just to split this into two services or something because as it is now they're all needed.

Also BTW I don't know why we have "getHarvester" and "getDkanHarvesterInstance" they seem totally redundant. I'll probably squash those into one method at some point.

[dafeder]

@stefan-korn do you plan to fix this? Thanks

[dafeder]

@stefan-korn bump :)

Make sure to rebase this if you pick it back up, as you need to switch to the new qlty tooling