compact: Simplify getting hashes in NewTreeWithState #1618
Conversation
Codecov Report
@@ Coverage Diff @@
## master #1618 +/- ##
=========================================
+ Coverage 66.89% 66.9% +0.01%
=========================================
Files 110 110
Lines 8946 8943 -3
=========================================
- Hits 5984 5983 -1
+ Misses 2358 2357 -1
+ Partials 604 603 -1
Continue to review full report at Codecov.
|
Codecov Report
@@ Coverage Diff @@
## master #1618 +/- ##
=========================================
+ Coverage 66.89% 66.9% +0.01%
=========================================
Files 110 110
Lines 8946 8943 -3
=========================================
- Hits 5984 5983 -1
+ Misses 2358 2357 -1
+ Partials 604 603 -1
Continue to review full report at Codecov.
|
| } | ||
| // Note: Right border nodes of compact.Range are ordered from root to leaves. | ||
| for i, j := 0, len(ids)-1; i < j; i, j = i+1, j-1 { | ||
| ids[i], ids[j] = ids[j], ids[i] |
There was a problem hiding this comment.
Just a thought but we could add the other ordering to Range if reversal is going to crop up a lot.
There was a problem hiding this comment.
Good thought, we'll see if/when another use-case appears. That wouldn't be a trivial change though. Range implementation has a very specific ordering: hashes slice contains 2 stacks of hashes - the front one is ordered bottom-to-top, the back one is ordered top-to-bottom. Stack ordering comes handy when 2 ranges are merged, so that the slice is mutated with minimal copies/reallocations.
* master: (94 commits) Complete TODO (#1632) fake_node_reader: Remove unused field (#1631) Parallelize VerifyMapLeavesResponse (#1615) Remove redundant root hash calculations (#1630) sequencer: Consolidate compact.Tree initialization (#1629) Remove unused NodeReader method (#1625) Fix bazel build (#1627) added/updated postgresql implementation of log_storage (#1571) Clean up compact.Tree tests (#1622) Remove old hash list format from compact.Tree (#1621) Mention protoc-gen-doc in README.md merkle: Add testonly package with golden hashes (#1620) compact: Simplify getting hashes in NewTreeWithState (#1618) compact: Implement Tree using Range (#1522) Guard verbose logging in merkle path code (#1604) Make MaskLeft of NodeID return an explicit copy. (#1612) Convert directly from Int.Bits() to NodeID (#1614) Couple of changes to make NodeIDs more frugal. (#1613) compact.Tree: Change semantic of adding leaves (#1592) MapHasher: Do not return error from HashLeaf (#1611) ...
This change factors out getting the list of node IDs needed to initialize
compact.Treeinto a separate exported function.
This eliminates
GetNodesFunccallback which makes client code simpler, and alsoallows a more direct initialization of
compact.Range.