Catchpoints: Small tweaks, mostly to comments #5195
Conversation
| id: childNodeID, | ||
| hashIndex: d[0], |
There was a problem hiding this comment.
FYI: This is because we write them in this order in all the nearby initializations.
Codecov Report
@@ Coverage Diff @@
## master #5195 +/- ##
==========================================
+ Coverage 53.55% 53.60% +0.05%
==========================================
Files 430 430
Lines 54085 54080 -5
==========================================
+ Hits 28966 28992 +26
+ Misses 22873 22848 -25
+ Partials 2246 2240 -6
... and 10 files with indirect coverage changes 📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
| @@ -63,7 +63,7 @@ type Trie struct { | |||
| root storedNodeIdentifier | |||
| nextNodeID storedNodeIdentifier | |||
| lastCommittedNodeID storedNodeIdentifier | |||
| cache *merkleTrieCache | |||
| cache merkleTrieCache | |||
There was a problem hiding this comment.
question: most of the changes in the PR that changes from mt.cache to &mt.cache are related with this line.
Is it because of memory management that we want to change from pointer to instance of cache?
There was a problem hiding this comment.
Yeah, it just didn't need to be a pointer, so I figured we'd avoid the allocation.
It was just a pointless allocation. The pointed to thing is never assigned into another merkleTrieCache, so it might as well be embedded. |
|
I just added a couple more very safe typo changes. If anyone wants me to back out the pointer change, I will. I'm just trying neaten things up before I dig into performance for state commitments. |
Existing tests pass, very few code changes.