ben-manes commented on a change in pull request #230:
URL: https://github.com/apache/solr/pull/230#discussion_r683821705
##########
File path: solr/core/src/java/org/apache/solr/search/CaffeineCache.java
##########
@@ -190,73 +191,72 @@ public V get(K key) {
return cache.getIfPresent(key);
}
- @Override
- public V computeIfAbsent(K key, IOFunction<? super K, ? extends V>
mappingFunction) throws IOException {
- /*
- This block will get deleted before commit, but I'm leaving it in for now
in case I have to go back to that apporach
+ private V computeAsync(K key, IOFunction<? super K, ? extends V>
mappingFunction) throws IOException {
+ // First do a lookup first: optimistic non-locking path.
+ // The cache does this for us on get(K, Function), but not putIfAbsent
+ V value = get(key);
+ if (value != null) {
+ return value;
+ }
Review comment:
Stats gets very funky with a Map, e.g. what stats are recorded for `V
replace(k, v)` and `boolean replace(k, old, new)`? In Guava we decided to not
record stats for the `asMap()` view and not cherry-pick. That made sense since
you could use `Cache.getIfPresent` and prefer a loading `get(k, loader)` over a
`putIfAbsent`.
In Caffeine, I decided that `compute` was valuable enough to break that
practice and record stats. I'd probably lean towards `asMap().get(key)` also
recording because it's been a surprise that `getIfPresent` should be used
instead, but we have Guava's history to justify keeping as is. Similarly
putIfAbsent sounds good, but it gets too confusing about the "correct" stats.
Thus the fallback is to tell users to not rely on ours and create their own to
get what they deem to be the right metrics. Ours are therefore best effort, but
fallible.
--
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...@solr.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org
