Commit to a next power table and additional commitments #273
Conversation
|
Posting this so we can discuss with respect to #151. |
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #273 +/- ##
==========================================
- Coverage 71.78% 70.11% -1.67%
==========================================
Files 35 35
Lines 2545 2617 +72
==========================================
+ Hits 1827 1835 +8
- Misses 588 651 +63
- Partials 130 131 +1
|
Stebalien
force-pushed
the
steb/root-power-table
branch
from
f4dcf71
to
ccce9d6
Compare
Stebalien
changed the title
|
Fuzz test failed on commit ccce9d6. To troubleshoot locally, download the seed corpus using GitHub CLI by running: gh run download 9218069385 -n testdataAleternatively, download directly from here. |
Stebalien
force-pushed
the
steb/root-power-table
branch
3 times, most recently
from
11ce69c
to
c46676d
Compare
|
The first commit contains all the changes, the last one just renames This isn't quite ready for merging yet, but I would like some feedback before I put the finishing touches (tests) on it. |
|
Fuzz test failed on commit c46676d. To troubleshoot locally, download the seed corpus using GitHub CLI by running: gh run download 9229851709 -n testdataAleternatively, download directly from here. |
Stebalien
force-pushed
the
steb/root-power-table
branch
2 times, most recently
from
dedcb40
to
c06302f
Compare
|
I think we're ready for a final review and merging here unless we want to wait for power table lookback to land first. |
Stebalien
force-pushed
the
steb/root-power-table
branch
2 times, most recently
from
00bab73
to
458718c
Compare
|
Ah, actually, let's land #264 today and then I can rebase/finish this one. |
Stebalien
force-pushed
the
steb/root-power-table
branch
from
458718c
to
1fb9a61
Compare
Stebalien
force-pushed
the
steb/root-power-table
branch
from
1fb9a61
to
3f8e683
Compare
In addition to committing to the power table for each tipset, we commit to the power table used to validate the next instance in the instance root (along with a merkle-tree of arbitrary commitments). fixes #257
That way we match tipsets. We could sign just the payloads in both cases but... alignment isn't _that_ important here. The only important part is that the first 3 words (32-byte) are aligned: 1. The first word has to be created inside the smart contract. The 32-byte alignment lets us AND the fixed prefix and the 8-byte instance number. 2. The second word is the commitment tree root and will likely be used by, e.g., snarks for validation. 3. The third word is the root of the tipset tree. But smart contracts aren't usually going to even _look_ at the power table CID. And if they do, the variable-size is going to be the least of their costs/complexity.
Stebalien
force-pushed
the
steb/root-power-table
branch
from
89c48a5
to
17a2552
Compare
In addition to committing to a power table in each tipset, agree on:
fixes #257