[jira] [Commented] (GEODE-10049) Redis tests should include the entire error response message rather than just the error type

["ASF subversion and git services (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/GEODE-10049?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17499195#comment-17499195 ] 

ASF subversion and git services commented on GEODE-10049:
---------------------------------------------------------

Commit d9ef0f15892c1993756e5fef5ef33db3431dc449 in geode's branch refs/heads/develop from Donal Evans
[ https://gitbox.apache.org/repos/asf?p=geode.git;h=d9ef0f1 ]

GEODE-10049: Check full Redis error message (#7362)

 - Replace all uses of hasMessageContaining() with hasMessage() in Redis
 tests that assert on the error message returned, with the exception of
 CLIENT, CLUSTER and COMMAND, which return a more complex error message
 - Return correct error messages for MSET and MSETNX with even
 number of arguments
 - Return correct error message for BITOP NOT when called with too many
 arguments
 - Simplify error message generation in UnknownExecutor
 - Remove tests for behaviour with no longer existent internal commands
 - For error responses that always use the same message (OOM, CROSSSLOT,
 BUSYKEY, WRONGPASS, WRONGTYPE, NOAUTH) do not take a String argument in
 RedisResponse
 - Remove prefix-specific error generating methods from Coder in favour
 of calling generic getErrorResponse()
 - Remove isInstanceOf assertions on exception type returned by Jedis,
 as it's something we have no control over
 - Replace hard-coded error messages with references to constants
 - Add missing error message assertions to a few tests
 - Format error assertions to be single-line where possible
 - Move error constants from SetEXExecutor to RedisConstants
 - Return correct error for non-integer expiration argument in SETEX and
PSETEX
 - Add tests for non-integer expiration argument for SETEX and PSETEX
 - Add tests for non-integer expiration argument for SET with EX and PX
options

Authored-by: Donal Evans <do...@vmware.com>

> Redis tests should include the entire error response message rather than just the error type
> --------------------------------------------------------------------------------------------
>
>                 Key: GEODE-10049
>                 URL: https://issues.apache.org/jira/browse/GEODE-10049
>             Project: Geode
>          Issue Type: Test
>          Components: redis
>    Affects Versions: 1.16.0
>            Reporter: Donal Evans
>            Assignee: Donal Evans
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: 1.16.0
>
>
> Currently many tests look for substrings of error messages, rather than the error message as a whole. This has in the past led to cases where the Geode for Redis Module's error messages have not precisely matched those of native Redis.
> For example, if the test is:
> {code:java}
> assertThatThrownBy(
>         () -> jedis.hsetnx(string_key, field, "something else"))
>             .isInstanceOf(JedisDataException.class)
>             .hasMessageContaining("WRONGTYPE");{code}
> instead we should probably look for the full error message that native Redis puts out:
> {code:java}
> .hasMessage("WRONGTYPE Operation against a key holding the wrong kind of value"){code}



--
This message was sent by Atlassian Jira
(v8.20.1#820001)