[
https://issues.apache.org/jira/browse/IMPALA-14174?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Csaba Ringhofer updated IMPALA-14174:
-------------------------------------
Description:
.cls and .self are often used interchangeably in subclasses of ImpalaTestSuite,
which can get confusing, e.g
https://gerrit.cloudera.org/#/c/22527/23/tests/metadata/test_event_processing_base.py
A reason is that many initialization logic happens in
ImpalaTestSuite/CustomClusterTestSuite's setup_class, which can only use class
methods, but some of these class methods are also used during per test method
initialization, where instance methods would be more suitable. An example is
cluster_setup()
https://github.com/apache/impala/blob/892b33f35d1cab01d91d9b5eaac079ec8ba236bb/tests/common/custom_cluster_test_suite.py#L268
class and instance variables as not differentiated in naming but have different
subclass semantics, e.g CustomClusterTestSuite.my_var = 1 won't be visible when
someone asks for ImpalaTestSuite.my_var
Note that the classic pytest answer would be "don't use any of those, use
fixtures", but turning some of these members (clients, cluster abstraction)
into fixtures would be a massive refactor and may make the code more noisy.
was:
.cls and .self are often used interchangeably in subclasses of ImpalaTestSuite,
which can get confusing, e.g
https://gerrit.cloudera.org/#/c/22527/23/tests/metadata/test_event_processing_base.py
A reason is that many initialization logic happens in
ImpalaTestSuite/CustomClusterTestSuite's setup_class, which can only use class
methods, but some of these class methods are also used during per test method
initialization, where instance methods would be more suitable. An example is
cluster_setup()
https://github.com/apache/impala/blob/892b33f35d1cab01d91d9b5eaac079ec8ba236bb/tests/common/custom_cluster_test_suite.py#L268
class and instance variables as not differentiated in naming but have different
subclass semantics, e.g CustomClusterTestSuite.my_var = 1 won't be visible when
someone asks for ImpalaTestSuite.my_var
> Clean up class vs instance handling in Impala EE tests
> ------------------------------------------------------
>
> Key: IMPALA-14174
> URL: https://issues.apache.org/jira/browse/IMPALA-14174
> Project: IMPALA
> Issue Type: Improvement
> Components: Test
> Reporter: Csaba Ringhofer
> Priority: Minor
>
> .cls and .self are often used interchangeably in subclasses of
> ImpalaTestSuite, which can get confusing, e.g
> https://gerrit.cloudera.org/#/c/22527/23/tests/metadata/test_event_processing_base.py
> A reason is that many initialization logic happens in
> ImpalaTestSuite/CustomClusterTestSuite's setup_class, which can only use
> class methods, but some of these class methods are also used during per test
> method initialization, where instance methods would be more suitable. An
> example is cluster_setup()
> https://github.com/apache/impala/blob/892b33f35d1cab01d91d9b5eaac079ec8ba236bb/tests/common/custom_cluster_test_suite.py#L268
> class and instance variables as not differentiated in naming but have
> different subclass semantics, e.g CustomClusterTestSuite.my_var = 1 won't be
> visible when someone asks for ImpalaTestSuite.my_var
> Note that the classic pytest answer would be "don't use any of those, use
> fixtures", but turning some of these members (clients, cluster abstraction)
> into fixtures would be a massive refactor and may make the code more noisy.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
