[weld-dev] WIP fix for https://issues.jboss.org/browse/WELD-1130

[Benjamin Confino]

Hello Martin. 

Thank you for the feedback. 

Moving the change to AbstractSessionBeanStore.get(BeanIdentifier) sounds 
like a good idea, and I am happy to change the config property name to 
anything you would like. What did you have in mind? 

In regards to point 2, I believe my customer's configuration will afford 
us some leeway. Providing that the caching still allows a call to get() 
once every two minuets I think we will not have a problem. As always 
though, the proof will be testing it, I was able to get the customer's 
test application working with the WIP fix I wrote so hopefully their live 
application will work too. 

If you can think of any strategy for writing an automated test let me 
know. Apart from that I expect I will have new commits based on your 
suggestion soom. 

Regards
Benjamin





From:   Martin Kouba <mkouba at redhat.com>
To:     Benjamin Confino <BENJAMIC at uk.ibm.com>, weld-dev at lists.jboss.org
Date:   18/04/2018 13:04
Subject:        Re: [weld-dev] WIP fix for 
https://issues.jboss.org/browse/WELD-1130



Hi Benjamin,

I'm sorry but the Weld team is burried under some high priority issues. 
I've quickly went through your proposal and have few comments.

1. I think it would be more convenient to modify 
org.jboss.weld.module.web.context.beanstore.http.AbstractSessionBeanStore.getAttribute(String) 

which is used for both the HTTP session and the conversation contexts

2. +1 for config property, we probably need a better name though ;-)

3. You're probably aware that this would not solve the problem 
completely. It should trigger the replication when a @SessionScoped bean 
instance is obtained from the context, e.g. when a client proxy method 
is invoked. However, there are some optimizations in Weld which prevent 
the context lookup (see also 
org.jboss.weld.bean.ContextualInstanceStrategy). Also the replication 
may happen before the state is actually changed - it really depends on 
replication impl details. Still, I agree that this would improve the 
usability.

4. +10 for tests. But I guess it would be tricky to write an automated 
test for this.

Martin

Dne 13.4.2018 v 21:33 Benjamin Confino napsal(a):
> Hello
> 
> I posed 
> 
https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_benjamin-2Dconfino_core_commit_35d055cd9a841cfbdf2bcb2aaa3e0d8d6116ec1cto&d=DwIC-g&c=jf_iaSHvJObTbx-siA1ZOg&r=hl8XiFr1UrTSgyXVc4KO_i51sM3Gyhlu1KZ4TkyseEQ&m=h44psnUUWe961Di4IEKfnHhsSfRzXKxVIpAMgp2r0LI&s=v_EHx3focK723Wv9TX7fVy4pYYq7or0XSJVFhVrezPo&e= 

> this lift list month but have not heard any feedback. When can I expect 
> to hear if this concept is good in principal? If the concept is good I 
> can begin investigating how to extend the idea to conversation scopes 
> and write a test.
> 
> Regards
> Benjamin
> Unless stated otherwise above:
> IBM United Kingdom Limited - Registered in England and Wales with number 

> 741598.
> Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 
3AU
> 
> 
> _______________________________________________
> weld-dev mailing list
> weld-dev at lists.jboss.org
> 
https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.jboss.org_mailman_listinfo_weld-2Ddev&d=DwIC-g&c=jf_iaSHvJObTbx-siA1ZOg&r=hl8XiFr1UrTSgyXVc4KO_i51sM3Gyhlu1KZ4TkyseEQ&m=h44psnUUWe961Di4IEKfnHhsSfRzXKxVIpAMgp2r0LI&s=zmOPLMTgosB3_X_zWg_rgSt6EgDlEZTmkAKbg6eRs-o&e=

> 




Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/weld-dev/attachments/20180418/de0c4031/attachment.html