Use original result when retry-https encounters an error
How to Test
Start with retry-https disable and find a page that redirects to a SSO page -> get a nice "Too many redirects" error.
Now enable retry-https and rescan -> you should still get the "Too many redirects" error, not a timeout or "tls oversized record received with length 20527"
Notes & Caveats
Currently, whenever any error is encountered and retry-https is enabled, the original, often still useful error is thrown away and instead you get either timeouts or "tls oversized record received with length 20527" when it fails, too. Especially with errors like "Too many redirects", which still can be useful given that the body and HTTP headers are still included, this retry behavior is infuriating and harmful at best, given that the retry error usually is super useless unless this port actually uses HTTPS.
Therefore, let's keep the original error when retry-https fails too, and only use retry-https' result when it actually succeeds.
Use original result when retry-https encounters an error
How to Test
Start with retry-https disable and find a page that redirects to a SSO page -> get a nice "Too many redirects" error. Now enable retry-https and rescan -> you should still get the "Too many redirects" error, not a timeout or "tls oversized record received with length 20527"
Notes & Caveats
Currently, whenever any error is encountered and retry-https is enabled, the original, often still useful error is thrown away and instead you get either timeouts or "tls oversized record received with length 20527" when it fails, too. Especially with errors like "Too many redirects", which still can be useful given that the body and HTTP headers are still included, this retry behavior is infuriating and harmful at best, given that the retry error usually is super useless unless this port actually uses HTTPS.
Therefore, let's keep the original error when retry-https fails too, and only use retry-https' result when it actually succeeds.