DOTCOM-25970 - in texts links lead to 404 (wrong link are rendered)

[avanishgandhi]

JIRA - https://jira.corp.adobe.com/browse/DOTCOM-25970?src=confmacro

Detailed Description:
................................
URL:https://theblog.adobe.com/five-ways-make-sure-distributors-love-benefits-brand-portals-manufacturer-distributor-relationships/

https://blog.adobe.com/five-ways-make-sure-distributors-love-benefits-brand-portals-manufacturer-distributor-relationships/

The link goes to a wrong URL - https://blog.adobe.com/en/2017/06/06/www.adobe.com/marketing/target/product-recommendations.html which seems to be common to this URL

................................
Steps to Reproduce:

1. click on digital asset management and product recommendations in text
   Expected Results:

Links go to new page
................................DOTCOM-25302
Actual Results:

404 error

[rofe]

The links are wrong in the content: they are missing the https:// prefix and are therefore (correctly) treated as relative URLs by the browser...

Interestingly, the original blog post has a class auto-replaced on the anchor tag, so I assume WP auto-corrects such incomplete URLs on the fly, but I wouldn't recommend doing that. Depending on how common they are, I think we should fix them directly in the content. For future posts, we could try to detect and auto-correct them in the importer.

cc @kptdobe @davidnuescheler

[rofe]

Fixed the 2 broken links in the content and added an enhancement request to the importer project. Closing.

[rofe]

Reopening after talking to @davidnuescheler: we will (also) fix them on the client side.

[trieloff]

@rofe should the detection of "likely links" be something done in the pipeline? e.g.

[Example](www.example.com)

becomes

<a href=www.example.com" data-likely-href="https://www.example.com">Example</a>

so that we can override the href with data-likely-href either in helix-markup.yaml or in a short client-side script?

[rofe]

should the detection of "likely links" be something done in the pipeline?

I think that would be a nice feature.

[rofe]

Although in theblog, we can be "dumb" and always assume absolute URLs, whereas in the pipeline, we'll need to be smarter and account for all kinds of link types.