Links on web-archived pages to arquivo.pt/wayback should not be rewritten

[dcgomes]

Consider a page archived from the live-web that had links to web-archived pages:

• https://memoriafcsh.wordpress.com/2017/08/01/centro-de-estudos-historicos-2000-2015/
  "A primeira versão preservada com o endereço fcsh.unl.pt/ceh é de 2000 e não tem alterações até 2006."
  links to:
• http://arquivo.pt/wayback/20000915125205/http://www.fcsh.unl.pt/ceh/

When the web-page becomes web-archived:
https://arquivo.pt/wayback/20180919142909/https://memoriafcsh.wordpress.com/2017/08/01/centro-de-estudos-historicos-2000-2015/
"A primeira versão preservada com o endereço fcsh.unl.pt/ceh é de 2000 e não tem alterações até 2006."
links to:
https://arquivo.pt/wayback/20180919142909mp_/http://arquivo.pt/wayback/20000915125205/http://www.fcsh.unl.pt/ceh/
which originates a "Not Archived" message because the Replay system rewrites the URL to and adds the prefix "https://arquivo.pt/wayback/20180919142909mp_/" to redirect the link to Arquivo.pt.

Wayback should detect when the URLs in the links targeted web-archived pages, don't add the prefix and keep the original URLs in web-archived pages when they link to "https://arquivo.pt/wayback/*". In this example, the URL should be kept as in the original live-web page:
https://arquivo.pt/wayback/20000915125205/http://www.fcsh.unl.pt/ceh/

[igobranco]

I think this could be done on Apache HTTPd site configuration.

[danielbicho]

The server redirect doesn't work, weird things start to happen:

I don't understand the pywb rewritting rules. will postpone this

[igobranco]

Can you put the apache httpd configuration on here so we can analyze it.

[danielbicho]

RewriteRule ^/wayback/.*/.arquivo.pt/wayback/(.)$ %{REQUEST_SCHEME}://%{SERVER_NAME}/wayback/$1 [PT]

It works if you open a new tab, or on an unframed context though

[dcgomes]

Check if newer version of PyWB fixes this issue.

[dcgomes]

Exceptions to rewrite of URLs was asked to Ilya.

[dcgomes]

Re-evaluate after Eros deploy

[VascoRatoFCCN]

I'll investigate implementing a fix on the replay page using pywb UI customization: no longer exists.

[arquivo-awp]

Test when we integrate new version of pywb.
@VascoRatoFCCN report issue to pywb project and close.