dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1380008658
##########
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##########
@@ -328,7 +298,100 @@ private void rehash(long lastNodeAddress) throws
IOException {
}
mask = newMask;
- entries = newEntries;
+ fstNodeAddress = newEntries;
+ copiedNodeAddress = newCopiedOffsets;
+ }
+
+ // hash code for a frozen node. this must precisely match the hash
computation of an unfrozen
+ // node!
+ private long hash(long node, long pos) throws IOException {
+ FST.BytesReader in = getBytesReader(node, pos);
Review Comment:
Actually I realized that nodeAddress can be retrieved from hashSlot, which
is just getting from the fstNodeAddress with the hash slot. It means we need an
extra get() (since the nodeAddress is already known by this point). But that
would also make it more consistent? Like passing an inconsistent value of
hashSlot and nodeAddress could be a problem. (We can get the nodeAddress based
on the hashSlot, but not the other way)
The end results is simpler
```
assert primaryTable.hash(hashSlot) == hash
: "mismatch frozenHash=" + primaryTable.hash(hashSlot) + " vs
hash=" + hash;
```
But yes the hash function would need an extra get()
```
private long hash(long hashSlot) throws IOException {
long nodeAddress = fstNodeAddress.get(hashSlot);
FST.BytesReader in = getBytesReader(nodeAddress, hashSlot);
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
