lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chris Hostetter <>
Subject Re: lucene-solr:master: disable query cache when this test uses threads, so seeds reliably reproduce
Date Mon, 28 Mar 2016 22:07:20 GMT

Mike, can you clarify why the query cache is problemtic here ... it's not 
really clear from the comment you added...

: +    if (useThreads) {
: +      // We must disable query cache otherwise test seed may not reproduce since different
: +      // threads may or may not get a cache hit or miss depending on order the JVM
: +      // schedules the threads:
: +      s.setQueryCache(null);
: +    }
: +

...regardless of wether any thread gets a cache hit or a cache miss, 
shouldn't the *results* of any arbitrary query as seen by the test 
thread always be the same?

I realize rob already removed all the threading code from this test in 
LUCENE-7136, i'm just trying to understand the general problem you were 
trying to fix here: how/why a cache hit/mass would have caused 
non-reproducible behavior in a threaded test.  

(unless something in that test was expecting/asserting a cache hit/miss, 
but i could find any evidence of that)


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message