[ 
https://issues.apache.org/jira/browse/MRESOLVER-464?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17807831#comment-17807831
 ] 

ASF GitHub Bot commented on MRESOLVER-464:
------------------------------------------

rmannibucau commented on PR #408:
URL: https://github.com/apache/maven-resolver/pull/408#issuecomment-1896294912

   > the fact HttpClient inits this value to Integer.MAX_VALUE and lowers it on 
first response
   
   just means each connection much await the first response, a semaphore does 
not make it right IMO, assuming each connection can get a single concurrent 
request and your pool is #10, how does your config helps? HttpClient does not 
enable to control connections nor to access http2 stream/frame data so looks 
like the workaround has the same bug than before to me.
   
   The system property does not queue but enables to avoid the first request 
(useless IMHO but similarly to the semaphore) and makes ` inits this value to 
Integer.MAX_VALUE` no more true.
   
   What you can do to scale more is to create more clients and limit one 
request per client, even if I don't like it much and I would just forbid http2 
for maven case, it would be saner than using a semaphone.
   
   ultimately, using the async flavor would probably also make it smoother 
since queueing will be simple to impl (IMHO - just a chain) but resolver is not 
greatly friendly with that option so maybe the multiple clients option will fit 
better short term?
   
   but from my window we kind of want to violate http2 so i'm really mixed and 
would just `version(HTTP_1_1)` since perf boost is not insane in practise.




> JDK transport bug
> -----------------
>
>                 Key: MRESOLVER-464
>                 URL: https://issues.apache.org/jira/browse/MRESOLVER-464
>             Project: Maven Resolver
>          Issue Type: Task
>          Components: Resolver
>            Reporter: Tamas Cservenak
>            Assignee: Tamas Cservenak
>            Priority: Major
>             Fix For: 2.0.0, 2.0.0-alpha-7
>
>
> JDK transport seems can be plagued by 
> https://bugs.openjdk.org/browse/JDK-8225647
> Try to do something about this.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to