lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hoss Man (JIRA)" <>
Subject [jira] [Commented] (SOLR-10235) fix last TestJdbcDataSource / mock issue with java9
Date Tue, 07 Mar 2017 17:36:37 GMT


Hoss Man commented on SOLR-10235:

bq.  I can add a note to the Javadocs why it is not a mock, but the purpose of this class
is identical to the other "mock" classes in this package (there is a mock JNDI Data Source
that behaves similar).

My point is that in _this_ test class (TestJdbcDataSource) the new {{MockDriver}} class is
the only "mock" class we explicitly define - all other code in this test uses Mockito based
mocks, so it would be easy for someone down the road to look at this class and think "this
MockDriver class looks like cruft we can replace with another one line Mockito call" ... basically
i'm just suggesting that virtually the same verbage from the {{testRetrieveFromDriverManager}}
method should be in  {{MockDriver}}'s javadocs.

bq. I disagree with that, sorry. As there can be more than one driver installed in the system,

AUGH! ... yoy are completely correct, i wasn't thinking holisticly about the DriverManager
as a whole ... agreed, ignore that suggestion completley.

> fix last TestJdbcDataSource / mock issue with java9
> ---------------------------------------------------
>                 Key: SOLR-10235
>                 URL:
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>            Reporter: Hoss Man
>              Labels: java9
>         Attachments: SOLR-10235.patch, SOLR-10235.patch, SOLR-10235.patch, SOLR-10235.patch
> The way TestJdbcDataSource was converted to use Mockito in SOLR-9966 still left one outstanding
test that was incompatible with Java9: {{testRetrieveFromDriverManager()}} 
> The way this test worked with mock classes was also sketchy, but under java9 (even with
Mockito) the attempt at using class names to resolve things in the standard SQL DriverManager
isn't viable.
> It seems like any easy fix is to create _real_ class (with a real/fixed classname) that
acts as a wrapper around a mockito "Driver" instance just for the purposes of checking that
the DriverManaer related code is working properly.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message