Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
Thank you for confirming. Looks like my team will have some work to do
next week!
---
Github user patricker commented on the issue:
https://github.com/apache/nifi/pull/3133
@colindean Your correct, patch releases are for very special circumstances.
I don't believe you have any option except to build this yourself until 1.9.0
gets released; and since 1.8.0 just came
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
ð
@patricker @mattyb149 How often and for what reasons does NiFi issue patch
releases? Looking through the release history, it seems pretty rare and saved
for regressions only. More
Github user patricker commented on the issue:
https://github.com/apache/nifi/pull/3133
@colindean Disregard. Git was misbehaving.
---
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
Very strange⦠it's there in the PR's commit and my local copy and Travis
built fine:
Github user patricker commented on the issue:
https://github.com/apache/nifi/pull/3133
@colindean I am getting compile errors. I had to add an import, `import
org.apache.nifi.components.PropertyValue;`, otherwise your method
`extractMillisWithInfinite` would not build.
This
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
Squash incomingâ¦
---
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
I'll squash once travis says ð
---
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
Test failure is on the Japanese build.
---
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
> an Idle connection was one that had been returned to the pool?
That's what I would think, but I couldn't seem to actually trigger it.
Reading through the API docs some more, I
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
Ah, I did that through checking the minIdle and maxIdle properties.
---
Github user patricker commented on the issue:
https://github.com/apache/nifi/pull/3133
My thought was you could look for the idle count and see if it was 0, 8,
etc... based on the config, and not worry about testing the timeouts for now.
---
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
> What if you exposed the number of active and idle connections in the
connection pool as properties on the DBCPConnectionPool? These are available by
calling getNumActive() and getNumIdle(). Or
Github user patricker commented on the issue:
https://github.com/apache/nifi/pull/3133
@colindean I know Matt responded right after I did before, what are your
thoughts on working on enabling unit tests by exposing the idle/active
connection counts?
---
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
Test failure now seems to be a timeout in another module:
```
[ERROR] Tests run: 5, Failures: 0, Errors: 1, Skipped: 0, Time elapsed:
30.287 s <<< FAILURE! - in
Github user patricker commented on the issue:
https://github.com/apache/nifi/pull/3133
@colindean I don't have a good answer yet, hoping I can get some input from
other developers.
But I was thinking about unit tests, and what you could do to help make
this code change unit
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
> What about making all of the new fields not required? `required(false)`.
Then, if a value is provided, you override, otherwise you leave it be. You
could call out the default values, or default
Github user colindean commented on the issue:
https://github.com/apache/nifi/pull/3133
> If adding new Properties, have you added .displayName in addition to
.name (programmatic access) for each of the new properties?
I added properties but I used a readable string for
18 matches
Mail list logo
