@nebulon Thanks for that info!

@Shaun-Snapp Check the logs of the FTP client you use. Sometimes it's set to use a certain protocol or security feature, but over the years that becomes obsolete and is replaced with a new one. You can also try creating a new connection and see if it picks up new settings.

@nebulon Works fine after updating. Thank you!

I wasn't aware of this feature, but it works perfectly. Thank you!

@jdaviescoates silly me

@girish the preview/download issue is fixed with the latest package version. Other two items are still pending though.

@jdaviescoates are you wanting to track individual files in a directory ? or when someone accesses the directory (listing) ?

FWIW, https://docs.cloudron.io , https://screenshots.cloudron.io are both surfer. For reasonable traffic sites, they work just fine. In the end, it very much depends on your use case. CDNs make sense if you have many static assets and a very very high traffic website. Google published some study a while ago that people go away if your page takes more than 3 seconds to load. So, that's a good metric to work with.

I haven't actually chimed in on this since posted I guess. Whoops... Yeah my ask would maybe be the ability to add redirects based on status codes. As was mentioned, for example, any 404 could route to index.html (or any file one wants to specify). Just gives more options. Like this in cloud front, for instance: [image: 1647634717369-screen-shot-2022-03-18-at-2.18.11-pm-resized.png]

@nebulon Perfect, thank you! @nebulon @girish thanks for the fast response on this!

@msbt I also forgot about that option actually, such a long time ago when I wrote all that code

more likely a browser hiccup..

@fbartels that is correct, admins can login to all apps always. This was the case from the get-go and now it is a bit hard to change this behavior since setups depend on this. But maybe worth reevaluating.

@nebulon said in Google Sign In for Surfer: so you could write small shell scripts for each site to avoid the risk of uploading to the wrong instance. Yes, I do the same. Thank you.

@luckow Thanks for introducing me to this site (siwecos.de)!

Confirmed!

Looking good, thanks for the fix!

Ok I added those test cases and made a new package release now. Please be aware that if you just use # in a path in the browser URL bar, without encoding it, the browser will treat this as a URL hash, not as part of the filepath. eg. foo/bar#baz.txt will end up as foo/bar only.