What happens to images depends on how you access LFGSS.
If you access via http://
Then you just get the images from the remote servers... i.e. Flickr, Facebook, Imgur or wherever the image is hosted.
In which case, I cannot help... they're nothing to do with me.
If you access LFGSS via https://
Then what happens depends on where the image comes from.
If it's from LFGSS, I just serve the image.
If it's from Facebook, Flickr, etc... then you get the image from them over their https:// connection.
If it's from any http:// source that doesn't have a https:// source, then I do get involved and do something strange. I proxy them... which means I have my server go get the copy (over http://), and then my servers sends you the copy over https://
This is to ensure that you get a green lock in the top of the screen and no errors about mixed content.
The issues that can occur with my forwarding images are:
1) I'm using a service called Heroku, and that might go down.
2) Heroku uses Amazon S3 cloud services, and that might go down.
3) In front of both of those is a service called Amazon Cloud Front, and that might go down.
Such occurrences should be rare, and will only affect http:// images requested over https:// and proxied through the https://sslcache.se/ domain.
All make sense?
In which case... I don't know why images suddenly stopped in the examples given.
The first 3 are all on Flickr, and Flickr offer https:// so you're now just getting the secure version of the image. But if the image is broken... it's 100% Flickr's doing.
And the ${secretKey} is a HMAC based on a shared secret and hash with the image URL. It's only there to prevent non-LFGSS sites from using sslcache to proxy images.
What sslcache does is:
1) Check the imageUrl is good (validated by the secretKey)
2) Fetch the image from imageURL
3) Serve the image
That's literally it, there's no logic to speak of... I just fetch the image and serve it.
Again, I only do this for http:// images that have no https:// source when you access via LFGSS over https://
And I can't explain the broken images above, as 90% of them in the examples shown come from remote servers and I've not touched the links at all.
So... I dunno what's happening there. Temporary glitch in the wider matrix?
What happens to images depends on how you access LFGSS.
If you access via http://
Then you just get the images from the remote servers... i.e. Flickr, Facebook, Imgur or wherever the image is hosted.
In which case, I cannot help... they're nothing to do with me.
If you access LFGSS via https://
Then what happens depends on where the image comes from.
If it's from LFGSS, I just serve the image.
If it's from Facebook, Flickr, etc... then you get the image from them over their https:// connection.
If it's from any http:// source that doesn't have a https:// source, then I do get involved and do something strange. I proxy them... which means I have my server go get the copy (over http://), and then my servers sends you the copy over https://
This is to ensure that you get a green lock in the top of the screen and no errors about mixed content.
The issues that can occur with my forwarding images are:
1) I'm using a service called Heroku, and that might go down.
2) Heroku uses Amazon S3 cloud services, and that might go down.
3) In front of both of those is a service called Amazon Cloud Front, and that might go down.
Such occurrences should be rare, and will only affect http:// images requested over https:// and proxied through the https://sslcache.se/ domain.
All make sense?
In which case... I don't know why images suddenly stopped in the examples given.
For example, in Jambons post ( https://www.lfgss.com/thread91194.html ) are these image links:
http://farm9.staticflickr.com/8421/7727779064_bd3f6be87b_c.jpg
http://farm9.staticflickr.com/8287/7727781106_fafd939d48_c.jpg
http://farm9.staticflickr.com/8288/7727782798_0794046612_c.jpg
http://img705.imageshack.us/img705/9503/systemsix2.jpg
Notice that they're all http:// links, so when you access LFGSS via https:// you would normally get a mixed content warning and won't see the images.
To ensure that you see the images, when you access LFGSS over https:// the image links are changed to:
https://farm9.staticflickr.com/8421/7727779064_bd3f6be87b_c.jpg
https://farm9.staticflickr.com/8287/7727781106_fafd939d48_c.jpg
https://farm9.staticflickr.com/8288/7727782798_0794046612_c.jpg
https://sslcache.se/a5bd740f0d18df8ab499166494da34e78c1ac1e0/687474703a2f2f696d673730352e696d616765736861636b2e75732f696d673730352f393530332f73797374656d736978322e6a7067
Notice that now they're all https:// links.
The first 3 are all on Flickr, and Flickr offer https:// so you're now just getting the secure version of the image. But if the image is broken... it's 100% Flickr's doing.
The last image is ImageShack.
The URL is:
http://img705.imageshack.us/img705/9503/systemsix2.jpg
But no image exists at:
https://img705.imageshack.us/img705/9503/systemsix2.jpg
So I will have to proxy it:
https://sslcache.se/a5bd740f0d18df8ab499166494da34e78c1ac1e0/687474703a2f2f696d673730352e696d616765736861636b2e75732f696d673730352f393530332f73797374656d736978322e6a7067
The sslcache URL is made up of two bits:
https://sslcache.se/${secretKey}/${imageUrl}
The ${imageUrl} is simply http://img705.imageshack.us/img705/9503/systemsix2.jpg hex encoded (so I can pass it around safely).
And the ${secretKey} is a HMAC based on a shared secret and hash with the image URL. It's only there to prevent non-LFGSS sites from using sslcache to proxy images.
What sslcache does is:
1) Check the imageUrl is good (validated by the secretKey)
2) Fetch the image from imageURL
3) Serve the image
That's literally it, there's no logic to speak of... I just fetch the image and serve it.
Again, I only do this for http:// images that have no https:// source when you access via LFGSS over https://
And I can't explain the broken images above, as 90% of them in the examples shown come from remote servers and I've not touched the links at all.
So... I dunno what's happening there. Temporary glitch in the wider matrix?