[
https://issues.apache.org/jira/browse/SOLR-16498?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17626726#comment-17626726
]
Chris M. Hostetter commented on SOLR-16498:
-------------------------------------------
Maybe i'm missing something, but this feels like a false premise...
{quote}We would prefer a non-dependency approach using [Java 11's new
HttpClient|https://docs.oracle.com/en/java/javase/11/docs/api/java.net.http/java/net/http/package-summary.html].
And we need to figure out how to get the URL *_from the SolrClient_* in a more
general way.
{quote}
If we're only concerned here with tests, then why does the URL need to come
_from the SolrClient_ ? If there is a SolrClient pointed at a URL in a test,
then that URL is coming from a (test) Jetty instance, so why can't we just ask
that jetty instance for it's URL directly?
Skimming the PR, i only see 2 tests being updated (are there others where this
issue is relevant?) and in both cases it seems straightforward to refactor the
tests so that instead of passing around a {{SolrClient}} just for the purposes
of getting a URL out of it, the test code could pass around the
{{JettySolrRunner}} to ask it for it's URL (or just pass around the URL)...
* {{TestRemoteStreaming}}
** instead of calling getSolrUrl({{{}getSolrClient(){}}}) just use
{{jetty.getBaseUrl()}}
** ({{{}jetty{}}} is a static variable in the super class)
* {{TestReplicationHandler}}
** refactor the private method {{adminClient(...)}} (and the two other private
methods that call it propogating their own args: {{reloadCore(...)}} and
{{watchCoreStartAt(...)}} ) so that instead of taking in a {{SolrClient}}
argument, it takes in a {{JettySolrRunner}} arg for the same purpose
** every existing place that calls these methods with a {{SolrClient}} also
has direct access to the corisponding {{JettySolrRunner}}
(or am i missunderstanding something fundamental about the goal here?)
> Tests need the URL from a SolrClient; don't require legacy clients
> ------------------------------------------------------------------
>
> Key: SOLR-16498
> URL: https://issues.apache.org/jira/browse/SOLR-16498
> Project: Solr
> Issue Type: Test
> Security Level: Public(Default Security Level. Issues are Public)
> Components: Tests
> Reporter: David Smiley
> Priority: Major
> Labels: newdev
> Time Spent: 1h 10m
> Remaining Estimate: 0h
>
> SOLR-16368 could not switch many tests away from some legacy/deprecated
> SolrClient classes because they were either getting the URL or Apache
> HttpClient instance out of the client. This issue aims to solve this so that
> our tests don't depend on a legacy client and don't depend on Apache
> HttpClient either. We would prefer a non-dependency approach using [Java
> 11's new
> HttpClient|https://docs.oracle.com/en/java/javase/11/docs/api/java.net.http/java/net/http/package-summary.html].
> And we need to figure out how to get the URL from the SolrClient in a more
> general way.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
