Skip to content

Releases: prep/beanstalk

Same as v1.4.6 but now with a proper tag

14 Sep 10:09
630ecac
Compare
Choose a tag to compare

URIs and unix domain sockets

08 Sep 14:38
630ecac
Compare
Choose a tag to compare
v.1.4.6

Update ValidURIs function to accept UDS (#25)

Export the Delete() method on Conn{}

26 Mar 14:22
3e8f9a9
Compare
Choose a tag to compare

This is a minor release that exports the Delete() method on Conn{}.

Add a configuration to avoid URI validation

14 Dec 10:24
5953ad7
Compare
Choose a tag to compare
v1.4.4

Add IgnoreURIValidation config flag to create consumer and producer w…

Don't disconnect on JOB_TOO_BIG errors

03 May 12:04
4f1a0d7
Compare
Choose a tag to compare
v1.4.3

Handle JOB_TOO_BIG errors more gracefully (#17)

Fix consumer bug and add job.Kick()

10 Mar 10:44
887d002
Compare
Choose a tag to compare
Implement kick-job (#13)

* feat: implement kick job

* tests: add tests cases for kick job

* feat: make kick job public

* tests: add test cases for kick

* just return err

* rename to kick and accept job as param

Remove span from the Consumer's reserveWithTimeout() call

24 Sep 14:37
776d8b9
Compare
Choose a tag to compare
v1.4.1

Remove tracing whenever the consumer reserves a job without timeout

Rewrite with simpler logic

01 Sep 08:44
1d0e426
Compare
Choose a tag to compare

This release introduces a rewrite of some of the logic in the code, most notably the consumer logic. With that comes a new API that no longer offers a Go channel on which to receive reserved jobs but instead offers a Receive() method on a consumer that spins up a fixed amount of workers. This approach makes the client connections less chatty and avoids reserving beanstalk jobs when there is no worker available to process it.

Don't fail if a connection to beanstalk could not be established immediately

13 Mar 18:26
85cf38e
Compare
Choose a tag to compare
v1.3.3

Create pool if there is atleast one valid server (#10)

Fix locking issue

11 Dec 16:06
3e23717
Compare
Choose a tag to compare

This release fixes an issue where a context timeout was set before it had acquired the connection lock. This resulted it i/o timeout errors. Now, a package-specific timeout can be configured that is started after the lock has been acquired.

Obviously, any timeout on the context that is coming from the outside are still subject to the old behaviour.