Bug - Links to old pages still show v5 components with no notice of it being the old version

[Venefilyn]

Describe the problem
Going to an older component that no longer exists, such as TextContent shows that it still exists on the main URL with no indication this is for the older v5.

https://www.patternfly.org/components/text/#textcontent

Expected behavior
I'd expect that this either

• redirects to the successor (Content),
• redirects to the old archived version,
• remains but with a display of it being outdated

Any other information?
We found this through the Patternfly website search

1. Go to search
2. Put in TextContent
3. See result for a component TextContent
4. Click and get the link mentioned above

See image, filed search issue separately: #4472

[andrew-ronaldson]

@nicolethoen and @edonehoo can we look at the Algolia stuff to find a fix

[Venefilyn]

@andrew-ronaldson I filed search issue in #4472. I might've been a bit unclear but with this issue I meant that the URL is showing www.patternfly.org instead of the archive for components that are no longer part of the current version, like these two are both the same despite the fact that the first one is old and not relevant anymore

• https://www.patternfly.org/components/text
• https://v5-archive.patternfly.org/components/text

With the www one if you go to that link and then go to other components you will see that the URL is still www but showing old v5 stuff. The version in top right will show the v5 as latest and v6 as alpha too so that doesn't help

[nicolethoen]

That is very interesting. We might need to remove some files from the S3 bucket (@kmcfaul).

I wonder if we should add the step of deleting all old files from the production S3 bucket before copying the files from the staging S3 bucket into the production bucket. That may be why this webpage is still appearing. I notice that the version switcher in your screenshots is saying it's displaying the 5.4.2 docs.

I'm not sure if there'd be negative side effects of trying that. (might be worth creating a backup of the current production bucket before we get too experimental). WDYT @kmcfaul? Does my theory make sense?