[
https://issues.apache.org/jira/browse/SOLR-8970?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hoss Man updated SOLR-8970:
---------------------------
Attachment: SOLR-8970.patch
bq. It would be nice if the keystore blob was included in the test jar...
Agreed, but I think that would be a lot trickier? ... i'm pretty sure the way
the keystore field is currently used in tests involves letting jetty load the
file itself via a path specified in the jetty.xml test config? in any case --
adding a sysprop to override seems like a good first step for now.
I'm attaching a patch with what i've got so far -- but i'm putting this on the
back burner until i verify it's working sanely with clientAuth ... AFAICT at
the moment, the clientAuth randomization isn't actually resulting in clientAuth
being required anywhere!
> SSLTestConfig behaves really stupid if keystore can't be found
> --------------------------------------------------------------
>
> Key: SOLR-8970
> URL: https://issues.apache.org/jira/browse/SOLR-8970
> Project: Solr
> Issue Type: Bug
> Reporter: Hoss Man
> Attachments: SOLR-8970.patch
>
>
> The SSLTestConfig constructor lets the call (notable SolrTestCaseJ4) tell it
> wether clientAuth should be used (note SolrTestCaseJ4 calls this boolean
> "trySslClientAuth") but it has a hardcoded assumption that the keystore file
> it can use (for both the keystore and the truststore) will exist at a fixed
> path in the solr install.
> when this works, it works fine - but if end users subclass/reuse
> SolrTestCaseJ4 in their own projects, they may do so in a way that prevents
> the SSLTestConfig keystore assumptions from being true, and yet they won't
> get any sort of clear error.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
