[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17749423#comment-17749423
]
Phil Steitz commented on POOL-411:
--
Fully fixed in 980aae56bf732178a9e02a3cb686f0104bffb6d3.
The problem
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741955#comment-17741955
]
Gary D. Gregory commented on POOL-411:
--
[~psteitz]
{quote}Any suggestions about how to do that in a
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741435#comment-17741435
]
Phil Steitz commented on POOL-411:
--
[~sebb] I added sleeps temporarily between the locks in
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741381#comment-17741381
]
Sebb commented on POOL-411:
---
The process of upgrading the lock has an inherent window when other threads can
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741331#comment-17741331
]
Phil Steitz commented on POOL-411:
--
[~sebb] I am not sure I understand your last comment. With the test
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741254#comment-17741254
]
Sebb commented on POOL-411:
---
I think it would be worth temporarily adding some sleeps/yields in the lock
upgrade
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741221#comment-17741221
]
Gary D. Gregory commented on POOL-411:
--
Committed to master. Because allocate is accessed in and out
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741220#comment-17741220
]
Phil Steitz commented on POOL-411:
--
[~ggregory] - looks great. Great idea to use computeIfAbsent and to
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741210#comment-17741210
]
Gary D. Gregory commented on POOL-411:
--
The following allows the test to pass locally for me in
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741196#comment-17741196
]
Gary D. Gregory commented on POOL-411:
--
I refactored the test and added
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741156#comment-17741156
]
Phil Steitz commented on POOL-411:
--
If I make the following changes to testConcurrentBorrowAnd clear, the
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17741106#comment-17741106
]
Phil Steitz commented on POOL-411:
--
[~sebb] - thanks for looking at this. You are right - the scenario I
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17740943#comment-17740943
]
Sebb commented on POOL-411:
---
To simplify testing, it might be worth temporarily adding a sleep where you
think
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17740941#comment-17740941
]
Sebb commented on POOL-411:
---
Surely it should not be possible to obtain a write lock whilst others are
holding
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17696566#comment-17696566
]
Richard Eckart de Castilho commented on POOL-411:
-
[~ggregory] I have only ever seen this
[
https://issues.apache.org/jira/browse/POOL-411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17696565#comment-17696565
]
Gary D. Gregory commented on POOL-411:
--
[~rec]
I added the null guard in {{deregister}}, please try
16 matches
Mail list logo