This is safer
Co-authored-by: Amaury Séchet <deadalnix@gmail.com>
Co-authored-by: Jason B. Cox <contact@jasonbcox.com>
Details
Details
- Reviewers
Mengerian schancel jasonbcox - Group Reviewers
Restricted Project - Commits
- rSTAGINGbcaa59bb2fbe: Hash the input of OP_CHECKDATASIG instead of expecting the user to do it.
rABCbcaa59bb2fbe: Hash the input of OP_CHECKDATASIG instead of expecting the user to do it.
Updated tests to reflect the change.
Diff Detail
Diff Detail
- Repository
- rABC Bitcoin ABC
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
| src/script/interpreter.cpp | ||
|---|---|---|
| 942 ↗ | (On Diff #4565) | Does ss.GetHash() return double-SHA256? That's how I wrote it in the Spec. Maybe this is crazy, but would it be possible to have the opcode use single-SHA256 for the hash function? Eg: https://bitco.in/forum/threads/gold-collapsing-bitcoin-up.16/page-1213#post-75916 |
| src/script/interpreter.cpp | ||
|---|---|---|
| 942 ↗ | (On Diff #4565) | Correct, this is double-sha256. |
| src/script/interpreter.cpp | ||
|---|---|---|
| 942 ↗ | (On Diff #4565) | OK, thanks. That confirms the spec is worded properly. You can ignore the rest of my above comment.. It was just an idea, but unclear if it's a benefit, and not worth changing this late. |
| test/functional/abc-checkdatasig-activation.py | ||
|---|---|---|
| 55 ↗ | (On Diff #4579) | Not sure about this fix. It looks like you intentionally NULLDUMMY the fail, and then OP_NOT it. However, that doesn't check that the signature was verified correctly? Are we asserting that the unit test does that instead? |
| test/functional/abc-checkdatasig-activation.py | ||
|---|---|---|
| 55 ↗ | (On Diff #4579) | Yeah, looks like it calls it in way that we can easily see that the opcode is called and should return FALSE. So test that the opcode is activated, but not whether it works properly for checking various signatures. |