hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Zoltan Haindrich <k...@rxd.hu>
Subject Re: Review Request 57343: HIVE-16127 Separate database initialization from actual query run in TestBeeLineDriver
Date Wed, 08 Mar 2017 16:43:03 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57343/#review168302
-----------------------------------------------------------



Hello Peter!

It's great to see that the qfile has a whole representative class now and the client aspect
is now more separated from it...it even makes reading it more easier!
I've found some minor issues - and added some notes here and there :)

cheers


itests/util/src/main/java/org/apache/hadoop/hive/cli/control/CoreBeeLineDriver.java
Lines 109 (patched)
<https://reviews.apache.org/r/57343/#comment240505>

    it would be great to use java.io.File (or Path or ?) api a bit more...and get rid of the
many string concatenations...



itests/util/src/main/java/org/apache/hadoop/hive/cli/control/CoreBeeLineDriver.java
Line 110 (original), 142 (patched)
<https://reviews.apache.org/r/57343/#comment240515>

    if the execution fails.. (parser error)..there is a very uncommunicative exception saying
null...could you add some description to this assertion: something like 'execution failed'
or something - it would be good to show the user that the test system worked as expected -
but the qfile have failed.



itests/util/src/main/java/org/apache/hadoop/hive/cli/control/CoreBeeLineDriver.java
Line 121 (original), 158 (patched)
<https://reviews.apache.org/r/57343/#comment240514>

    it would be great to give some more info about the failure...not sure what info is available
at this point - but I really miss the pointer to 'hive.log'...
    (possibly follow-up)



itests/util/src/main/java/org/apache/hadoop/hive/cli/control/CoreBeeLineDriver.java
Lines 165 (patched)
<https://reviews.apache.org/r/57343/#comment240516>

    i think this could be done with try-with-resources



itests/util/src/main/java/org/apache/hive/beeline/qfile/QFile.java
Lines 43-49 (patched)
<https://reviews.apache.org/r/57343/#comment240512>

    it would be great to have a bunch of File object references here...these strings always
cause troubles - here and there (possibly in a followup)



itests/util/src/main/java/org/apache/hive/beeline/qfile/QFile.java
Lines 100-103 (patched)
<https://reviews.apache.org/r/57343/#comment240517>

    instead of copyToDirectory...I would recommend to use the expected - I know they should
be the same... (or possibly there is some hidden magic which I'm not aware of :)



itests/util/src/main/java/org/apache/hive/beeline/qfile/QFile.java
Lines 105 (patched)
<https://reviews.apache.org/r/57343/#comment240518>

    I think it would be better to throw this execption out...the user expects it to be overwritten
- if this is the only thing that fails - it may cover-up to a non-updateable file like dir
and file owned by root by a mistake...



itests/util/src/main/java/org/apache/hive/beeline/qfile/QFile.java
Lines 266 (patched)
<https://reviews.apache.org/r/57343/#comment240519>

    I don't really see why staticfilterset is needed...and it's field - it seems to me that
they are used at the same time...



itests/util/src/main/java/org/apache/hive/beeline/qfile/QFileBeeLineClient.java
Lines 70 (patched)
<https://reviews.apache.org/r/57343/#comment240520>

    i don't think this should be asserted...if you run your jvm without -ea this method is
not even invoked



itests/util/src/main/java/org/apache/hive/beeline/qfile/QFileBeeLineClient.java
Lines 82 (patched)
<https://reviews.apache.org/r/57343/#comment240521>

    same assert related thing here


- Zoltan Haindrich


On March 8, 2017, 3:03 p.m., Peter Vary wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57343/
> -----------------------------------------------------------
> 
> (Updated March 8, 2017, 3:03 p.m.)
> 
> 
> Review request for hive, Zoltan Haindrich, Naveen Gangam, Sergio Pena, Vihang Karajgaonkar,
and Barna Zsombor Klara.
> 
> 
> Bugs: HIVE-16127
>     https://issues.apache.org/jira/browse/HIVE-16127
> 
> 
> Repository: hive-git
> 
> 
> Description
> -------
> 
> Refactored the QFileClient:
> - Moved to itest/util
> - Separated QFile specific code parts (file path, and filtering)
> - Separated BeeLineClient specific code (execution of the queries, commands)
> - Created factories for QFile, and Client, so after initialization they can be reused
during multiple tests
> - Separated init script run from actual test run, so multiple tests could be run against
a single MiniHS2 instance
> - Removed unused property
> 
> 
> Diffs
> -----
> 
>   beeline/src/java/org/apache/hive/beeline/util/QFileClient.java d306b7f 
>   itests/src/test/resources/testconfiguration.properties b01ebd8 
>   itests/util/src/main/java/org/apache/hadoop/hive/cli/control/CoreBeeLineDriver.java
aba1fde 
>   itests/util/src/main/java/org/apache/hive/beeline/qfile/QFile.java PRE-CREATION 
>   itests/util/src/main/java/org/apache/hive/beeline/qfile/QFileBeeLineClient.java PRE-CREATION

>   itests/util/src/main/java/org/apache/hive/beeline/qfile/package-info.java PRE-CREATION

>   ql/src/test/results/clientpositive/beeline/drop_with_concurrency.q.out PRE-CREATION

> 
> 
> Diff: https://reviews.apache.org/r/57343/diff/2/
> 
> 
> Testing
> -------
> 
> Used the previosly generated test to validate that the result is not changed.
> Added one more test to check the separation is done correctly.
> 
> 
> Thanks,
> 
> Peter Vary
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message