Open Akirathan opened 8 months ago
A duplicate of https://github.com/enso-org/enso/issues/6691 I think.
I think the scope is different. This one only implies the change of hashing algorithm in the text/applyEdit
request.
Also @Akirathan, if the goal is just to remove the bouncycastle dependency, can the version calclulator use the same logic with the MessageDigest
from the standard library as FileSystem
does? Without changing the algo
https://github.com/enso-org/enso/blob/689c8f7c3c9bb745d8d65e35384a22f084fd2160/engine/language-server/src/main/scala/org/enso/languageserver/filemanager/FileSystem.scala#L234-L248
Ok, I see, you already did it in #8664
Ok, I see, you already did it in #8664
Yes. Removing the org.bouncycastle
dependency is a related, but different task. And it is done in #8664.
I think the scope is different. This one only implies the change of hashing algorithm in the
text/applyEdit
request.
Fine, but there is a problem (at least currently) when we have to support both, old and new, GUI. They can't use different algorithms. I would say this would have to wait until old GUI is completely dropped and still will be an incompatible change.
this would have to wait until old GUI is completely dropped
Or we can fix the old GUI to use SHA-1. I assume downgrading the algorithm is going to be easy.
On hold until GUI1 is retired then we will switch to SHA1 in GUI2.
In
text-buffer
, SHA3-224 is used for calculating the checksum of a text content - https://github.com/enso-org/enso/blob/689c8f7c3c9bb745d8d65e35384a22f084fd2160/lib/scala/text-buffer/src/main/scala/org/enso/text/Sha3_224VersionCalculator.scala#L11According to the original author, @4e6, there is no reason to use specifically SHA3-224, so we can just use the simpler SHA-1.
Note that this change also requires modifications to some gui parts.
Related to:
8362
8553
8664