You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@lucene.apache.org by "Erick Erickson (Jira)" <ji...@apache.org> on 2020/07/02 14:53:00 UTC
[jira] [Comment Edited] (SOLR-13308) Move
SystemPropertiesRestoreRule to SolrTestCase
[ https://issues.apache.org/jira/browse/SOLR-13308?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17150271#comment-17150271 ]
Erick Erickson edited comment on SOLR-13308 at 7/2/20, 2:52 PM:
----------------------------------------------------------------
[~hossman@fucit.org] It looks like you already did something like this, but it's specific to RevertDefaultThreadHandlerRule. WDYT about taking that restriction away? Mostly I'm curious about why you made it so specific, do you see problems with making it more general or is this simply being cautious?
I'm a little nervous about RestoreRule affecting all tests. Would it make more sense to add a @BeforeClass that made a list of all the sysvars and then an @AfterClass that failed if the lists weren't identical? I don't particularly like adding that overhead, but it does seem safer.
[~hossman] think I used the wrong link to ping you before.
was (Author: erickerickson):
[~hossman@fucit.org] It looks like you already did something like this, but it's specific to RevertDefaultThreadHandlerRule. WDYT about taking that restriction away? Mostly I'm curious about why you made it so specific, do you see problems with making it more general or is this simply being cautious?
I'm a little nervous about RestoreRule affecting all tests. Would it make more sense to add a @BeforeClass that made a list of all the sysvars and then an @AfterClass that failed if the lists weren't identical? I don't particularly like adding that overhead, but it does seem safer.
> Move SystemPropertiesRestoreRule to SolrTestCase
> ------------------------------------------------
>
> Key: SOLR-13308
> URL: https://issues.apache.org/jira/browse/SOLR-13308
> Project: Solr
> Issue Type: Improvement
> Affects Versions: 8.1, master (9.0)
> Reporter: Erick Erickson
> Assignee: Erick Erickson
> Priority: Major
>
> [~krisden] just traced flaky HDFS tests to not clearing sysprops, see the discussion at SOLR-13297.
> Part of SOLR-13268 is to derive all Solr test classes from SolrTestCase.
> Once that's pushed, does it make sense to put that rule in SolrTestCase so it gets performed automagically so this kind of error doesn't creep back in?
> I don't know the details of how that rule is implemented or whether this makes sense at that level but wanted to discuss.
> I've assigned it to myself, but that's just so I don't lose track of it, but anyone else who wants to pick it up please feel free.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org