fix: sha256() for Slices and rework its description
#1936
Conversation
Also made the passed Slices use the same function, otherwise they'd still have all the same problems we've fixed before for Strings — the benchmarks didn't show those issues.
|
Looks like this should have a different linked issue. This is a fix of a fix. Can you please add an example that fails into a separate issue? |
Opened and linked: #1937 |
|
Is it still possible to recover the previous behavior? I.e. if the user want to hash just the slice data, without any traversals? |
novusnota
changed the title
sha256() as gas-expensive and rework its descriptionsha256() as gas-expensive, fix it for Slices and rework its description
If the Slice would not contain any references the results are the same as before. If the user still wants to hash just the data even if the slice contains some refs, in the "Before Tact 1.6" note I've said that TVM instructions that perform SHA-256 hashing all operate on data only. With that they can simply introduce their own helper Or do we want to add a new function to the stdlib? |
anton-trunov
changed the title
sha256() as gas-expensive, fix it for Slices and rework its descriptionsha256() for Slices and rework its description
I think we should. |
Sure thing, I've opened a separate issue for this — #1950. Will be resolved in a separate PR soon, let's finish this one first |
|
merge conflicts |
Resolved, please take a look |
Also made the passed Slices use the same function instead of the
string_hash()from FunC's stdlib, otherwise they'd still have all the same problems we've fixed before for Strings in #1626. Added positive and negative tests, showing that previous implementation for Slices was incorrect on Slices larger than 127 bytes.Issue
Closes #1879.
Closes #1937.
Checklist