Sorry about cross posting, but this will impact all developers and I wanted to give you all a heads-up.

HADOOP-9112<https://issues.apache.org/jira/browse/HADOOP-9112> was just checked it.  This means that the pre commit build will now give a –1 for any patch with junit tests that do not include a timeout option.  See http://junit.sourceforge.net/javadoc/org/junit/Test.html for more info on that.  This is to avoid surefire timing out junit when it gets stuck and not giving any real feedback on which test failed.

--Bobby

thanks, just seen and commented on this.

IF we're going to have test timeouts, we need a good recommended default
value for all tests except the extra slow ones.

Or
1. we just use our own JUnit fork that sets up a better default value than
0. I don't know how Ken would react to that.
2. we get maven sure to support a programmable (default) timeout for every
test case. This is probably easier, and would work for all existing tests
too.

Has anyone asked the Maven team about options here?

On 20 February 2013 16:46, Robert Evans <[EMAIL PROTECTED]> wrote:

> Sorry about cross posting, but this will impact all developers and I
> wanted to give you all a heads-up.
>
> HADOOP-9112<https://issues.apache.org/jira/browse/HADOOP-9112> was just
> checked it.  This means that the pre commit build will now give a –1 for
> any patch with junit tests that do not include a timeout option.  See
> http://junit.sourceforge.net/javadoc/org/junit/Test.html for more info on
> that.  This is to avoid surefire timing out junit when it gets stuck and
> not giving any real feedback on which test failed.
>
> --Bobby
>

Just wanted to give everyone a heads up. Due to this change, I have introduced a bug.

It's returning opposite value of what it's supposed to.
So, +1 of all tests you change brings up a -1 overall and vice versa.

I have reopened HADOOP-9112 and uploaded the patch to reverse to reverse to the correct return values. Once a committer goes ahead with, it will be corrected.

Sorry about the inconvenience.

Thanks,
Surenkumar Nihalani.

On Feb 21, 2013, at 5:28 AM, Steve Loughran <[EMAIL PROTECTED]> wrote:

> thanks, just seen and commented on this.
>
> IF we're going to have test timeouts, we need a good recommended default
> value for all tests except the extra slow ones.
>
> Or
> 1. we just use our own JUnit fork that sets up a better default value than
> 0. I don't know how Ken would react to that.
> 2. we get maven sure to support a programmable (default) timeout for every
> test case. This is probably easier, and would work for all existing tests
> too.
>
> Has anyone asked the Maven team about options here?
>
> On 20 February 2013 16:46, Robert Evans <[EMAIL PROTECTED]> wrote:
>
>> Sorry about cross posting, but this will impact all developers and I
>> wanted to give you all a heads-up.
>>
>> HADOOP-9112<https://issues.apache.org/jira/browse/HADOOP-9112> was just
>> checked it.  This means that the pre commit build will now give a –1 for
>> any patch with junit tests that do not include a timeout option.  See
>> http://junit.sourceforge.net/javadoc/org/junit/Test.html for more info on
>> that.  This is to avoid surefire timing out junit when it gets stuck and
>> not giving any real feedback on which test failed.
>>
>> --Bobby
>>