Re:Git 3.0's upcoming SHA-256 default will be a costly mistake
by crispr245 on Thursday October 01, 2026 @11:24PM
(#49929540)
Claude, make the hash use SHA-256 rather than SHA-1. No errors plsss.
Besides the possible implementation/deployment issues they will or will not face with this update, I can empathize with the idea that of not wanting to have a possible vector of attack in your system. Particularly today with AI being able to find novel exploits, I could see a future where a vulnerable hashing system leads to a malicious injection attack.
The author argues that "If I wanted to get untrusted code into Android, it's so much simpler to bribe or convince the maintainer of a popular downstream project" which is a really a red herring in this matter since that is literally a completely different issue that obviously no software update could ever fix.
Nonetheless I do agree with him in regards of how complicated and messy this whole process will be. Crypto migrations have been historically difficult, expensive and overall ugly, but not impossible...
Besides the possible implementation/deployment issues they will or will not face with this update, I can empathize with the idea that of not wanting to have a possible vector of attack in your system. Particularly today with AI being able to find novel exploits, I could see a future where a vulnerable hashing system leads to a malicious injection attack.
The author argues that "If I wanted to get untrusted code into Android, it's so much simpler to bribe or convince the maintainer of a popular downstream project" which is a really a red herring in this matter since that is literally a completely different issue that obviously no software update could ever fix.
Nonetheless I do agree with him in regards of how complicated and messy this whole process will be. Crypto migrations have been historically difficult, expensive and overall ugly, but not impossible...
https://nvlpubs.nist.gov/nistpubs/gcr/2018/NIST.GCR.18-017.p... page 58 for instance.