[
https://issues.apache.org/jira/browse/BEAM-1795?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15940756#comment-15940756
]
Daniel Halperin commented on BEAM-1795:
---------------------------------------
It's a question of:
- upgrade fast & get perf improvements / new bug fixes for free vs
- upgrade slow & potentially complicate users' lives by introducing some change
that breaks them
My intuition is to upgrade dependencies more slowly unless I need something (a
known bugfix or new feature). I certainly would rather not be a beta tester for
other projects ;). It also depends on the dependency -- core Java libs and
Maven plugins are much less likely to introduce backwards-incompatible changes
than some other libs.
I also suspect there is some aspect of language-dependence to this. Python has
fewer tools to integrate incompatible binaries (e.g., no shading AFAIK) -- so
it may make sense to upgrade faster assuming that's the standard Python
community behavior.
> Upgrade google-cloud-bigquery to 0.23.0
> ---------------------------------------
>
> Key: BEAM-1795
> URL: https://issues.apache.org/jira/browse/BEAM-1795
> Project: Beam
> Issue Type: Bug
> Components: sdk-py
> Reporter: Ahmet Altay
> Assignee: Mark Liu
>
> Should we upgrade this?
> https://pypi.python.org/pypi/google-cloud-bigquery/0.23.0
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
