[
https://issues.apache.org/jira/browse/BIGTOP-1315?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
jay vyas updated BIGTOP-1315:
-----------------------------
Description:
The pig tests which we ship are only running the TestPigTest and
TestGruntParser tests.
As usual, I'll make my trademark statement :)
1) Is all the indirection of including a jar file maintained externally really
worth it for two simple tests, neither of which are customizable, and both of
which run on very small data sets, built for a local machine only ? We can
easily maintain our own Itest based groovy tests in the Style of BIGTOP-1222.
Would be easier for others to use and adopt.
The second test "TestGruntParser" doesnt really seem like it should even run
inside of bigtop, should it?
2) If we still do want to keep using the artifacts from pigsmoke, for now we
will should to upgrade to pigsmoke 0.12.1.
My personal opinion (if you havent already guessed...) is that I think pig's
definition of a "smoke" test isnt quite the same as ours (unless im missing
something), so id like to think some more about (1), as a possible option /
alternative. :)
was:
The pig tests which we ship are only running the TestPigTest and
TestGruntParser tests.
As usual, I'll make my trademark statement :)
1) Is all the indirection of including a jar file maintained externally really
worth it for two simple tests, neither of which are customizable, and both of
which run on very small data sets, built for a local machine only ? We can
easily maintain our own Itest based groovy tests in the Style of BIGTOP-1222.
Would be easier for others to use and adopt.
The second test "TestGruntParser" doesnt really seem like it should even run
inside of bigtop, should it?
2) If we still do want to keep using the artifacts from pigsmoke, for now we
will need to upgrade to pigsmoke 0.12.1.
My personal opinion (if you havent already guessed...) is that I think pig's
definition of a "smoke" test isnt quite the same as ours (unless im missing
something), so id like to think some more about (1), as a possible option /
alternative. :)
> Pig smoke tests: Refactor ?
> ----------------------------
>
> Key: BIGTOP-1315
> URL: https://issues.apache.org/jira/browse/BIGTOP-1315
> Project: Bigtop
> Issue Type: Bug
> Components: Tests
> Reporter: jay vyas
> Assignee: jay vyas
> Fix For: 0.9.0
>
>
> The pig tests which we ship are only running the TestPigTest and
> TestGruntParser tests.
> As usual, I'll make my trademark statement :)
> 1) Is all the indirection of including a jar file maintained externally
> really worth it for two simple tests, neither of which are customizable, and
> both of which run on very small data sets, built for a local machine only ?
> We can easily maintain our own Itest based groovy tests in the Style of
> BIGTOP-1222. Would be easier for others to use and adopt.
> The second test "TestGruntParser" doesnt really seem like it should even run
> inside of bigtop, should it?
> 2) If we still do want to keep using the artifacts from pigsmoke, for now we
> will should to upgrade to pigsmoke 0.12.1.
> My personal opinion (if you havent already guessed...) is that I think pig's
> definition of a "smoke" test isnt quite the same as ours (unless im missing
> something), so id like to think some more about (1), as a possible option /
> alternative. :)
--
This message was sent by Atlassian JIRA
(v6.2#6252)
