mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-966267268
Well cool! You've got it, @dsmiley. I slightly improved the test. When it's
a time to commit it?
--
This is an automated message from the Apache Git Service.
To respond to the me
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-965761293
> I don't really follow but maybe you can show or explain further?
Anyway. I tried. it didn't worked out. I'd be happy to have a look on your
approach.
--
This is an auto
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-964836909
> In at least daemon() -- no; it misses the point.
Yep. I'm missing the point. 1) Why that daemon is shut down right there? Why
it's done asynchronously w/o awaitTerm?
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-964582922
we have (at least) two tread pool usages
https://github.com/mkhludnev/solr/blob/SOLR-15635-clear_SRI_twice/solr/solrj/src/java/org/apache/solr/client/solrj/io/stream/DaemonStream.ja
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-963105309
Sigh..
It seems `SRI.closed` can't be introduced. If a request thread spins off
`MDCAwareThreadPool` it's not obvious who should close `SRI`. In master branch,
it's closed twic
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-962421732
rolled back `SRI.addCloseHookStrict()` pushed more accurate version:
1. SRI.InheritableThreadlocal clears SRI but doesn't close SRI
2. introduced `SRI.closed`. it fails onl
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-961963576
> the problem here was that we end up calling close too many times, right?
Exactly `SRI.closeHook()` is invoked twice, one from the request thread, the
second one due to cleanin
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-961963576
> the problem here was that we end up calling close too many times, right?
Exactly `SRI.closeHook()` is invoked twice, one from the request thread, the
second one due to cleanin
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-961963576
> the problem here was that we end up calling close too many times, right?
Exactly `SRI.closeHook()` is invoked twice, one from the request thread, the
second one due to cleanin
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-959942709
@madrob, honestly, I hardly follow.
> we don't want to change things to strict. .. then we log it and ask the
person to create an issue
we do logging when refcount==-1, but
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-959942709
@madrob, honestly, I hardly follow.
> we don't want to change things to strict. .. then we log it and ask the
person to create an issue
we do logging when refcount==-1, but
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-959942709
@madrob, honestly, I hardly follow.
> we don't want to change things to strict. .. then we log it and ask the
person to create an issue
we do logging when refcount==-1, but
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-959942709
@madrob, honestly, I hardly follow.
> we don't want to change things to strict. .. then we log it and ask the
person to create an issue
we do logging when refcount==-1, but
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-955586812
I don't like `SRI.addCloseHookStrict()` too, @dsmiley. As another band-aid I
can put that atomic boolean wrapper into `JoinQuery` only. WDYT, @madrob ?
The question closer t
mkhludnev commented on pull request #376:
URL: https://github.com/apache/solr/pull/376#issuecomment-954160826
Hi, @dsmiley @NazerkeBS!
You review is quite appreciated!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and u
15 matches
Mail list logo
