Lifecycle refactor #190
Lifecycle refactor #190
Conversation
| configPath: configPath, | ||
| version: version, | ||
| loggingOpts: loggingOpts, | ||
| statusChan: make(chan *Status, 10), |
There was a problem hiding this comment.
Any reason we chose 10 here?
There was a problem hiding this comment.
It was an arbitrary selection related to the select statement below. I wanted to give a wide enough buffer that if something was listening to the status channel, it minimized the risk of missing something if it wasn't consuming fast enough.
| version: version, | ||
| loggingOpts: loggingOpts, | ||
| statusChan: make(chan *Status, 10), | ||
| wg: &sync.WaitGroup{}, |
There was a problem hiding this comment.
nit: usually wait groups are not pointers unless passed into functions. I have no idea why honestly more of just an observation you don't have to change it.
| select { | ||
| case c.statusChan <- status: | ||
| case c.statusChan <- &Status{running, err}: |
There was a problem hiding this comment.
Why is this in a select? Is it just so we don't block sending? Seems like we could miss a send if the channel is full. Maybe that's ok?
There was a problem hiding this comment.
I went with this pattern, because I didn't think there were any guarantees that whoever is using this struct is listening to the status channel. So I didn't want to lock everything up on a full channel, in the scenario where there is no consumer.
pkg/collector/test/invalid.yaml
Outdated
|
|
||
| service: | ||
| pipelines: | ||
| logs: | ||
| receivers: [filelog] | ||
| exporters: [logging] | ||
| exporters: [nop] |
There was a problem hiding this comment.
nit newline (you can thank Duri for me noticing missing newlines on files)
pkg/collector/test/valid.yaml
Outdated
|
|
||
| service: | ||
| pipelines: | ||
| logs: | ||
| receivers: [filelog] | ||
| exporters: [logging] | ||
| exporters: [nop] |
|
Tests detected a race condition, and lots of this in ci-check |
Proposed Change
Checklist