Re:JPEG XL would be Turing-complete without the 1024×1024 pi...
by jerryX on Monday June 21, 2021 @05:08AM
(#27577494)
What virus can be implemented in that?
[ Parent
| on HN ]
Re:JPEG XL would be Turing-complete without the 1024×1024 pi...
by jonsneyers on Monday June 21, 2021 @06:01AM
(#27577736)
The only 'virus' you could implement would be just a weird-looking image.
Re:JPEG XL would be Turing-complete without the 1024×1024 pi...
by slver on Monday June 21, 2021 @07:03AM
(#27578044)
Imagine I use a timing attack to visually encode some information about you in an image and trick you into sending a screenshot somewhere.
Very quick take, but systems are very interconnected these days, so sandboxes tend to leak in very unintentional ways.
Re:JPEG XL would be Turing-complete without the 1024×1024 pi...
by KMag on Monday June 21, 2021 @07:27AM
(#27578153)
How do you read a timer from within the JPEG XL bitstream? I believe the decoding process is deterministic.
Re:JPEG XL would be Turing-complete without the 1024×1024 pi...
by slver on Monday June 21, 2021 @07:34AM
(#27578190)
It's a challenge, but I'd probably look into any decoders that parallelize the decoding and see if I can abuse this.
Re:JPEG XL would be Turing-complete without the 1024×1024 pi...
by wizzwizz4 on Monday June 21, 2021 @07:38AM
(#27578221)
You'd need to exploit an implementation bug, or have the user be streaming or something (so some squares appear before others as the output of your timing attack).
Re:JPEG XL would be Turing-complete without the 1024×1024 pi...
by tsegratis on Monday June 21, 2021 @07:46AM
(#27578270)
Just wait, its only a short while til JPEG contains the pdf spec and a javascript interpreter
Re:JPEG XL would be Turing-complete without the 1024×1024 pi...
by jonsneyers on Monday June 21, 2021 @07:06AM
(#27578058)
Yes, that is true, but it isn't particularly related to JPEG XL specifically.