Re: Yes/No newbie question on contributing

2016-07-26 Thread Vinod Kumar Vavilapalli
The short answer is that it is expected to pass without any errors.

On branch-2.x, that command passes cleanly without any errors though it takes 
north of 10 minutes. Note that I run it with -DskipTests - you don’t want to 
wait for all the unit tests to run, that’ll take too much time. I expect trunk 
to be the same too.

Which branch are you running this against? What errors are you seeing? If it is 
unit-tests you are talking about, you can instead run with skipTests, run only 
specific tests or all tests in the module you are touching, make sure they pass 
and then let Jenkins infrastructure run the remaining tests when you submit the 
patch.

+Vinod

> On Jul 26, 2016, at 11:41 AM, Martin Rosse  wrote:
> 
> Hi,
> 
> In the How To Contribute doc, it says:
> 
> "Try getting the project to build and test locally before writing code"
> 
> So, just to be 100% certain before I keep troubleshooting things, this
> means I should be able to run
> 
> mvn clean install -Pdist -Dtar
> 
> without getting any failures or errors at all...none...zero, right?
> 
> I am surprised at how long this is taking as errors keep cropping up.
> Should I just expect it to really take many hours (already at 10+) to work
> through these issues? I am setting up a dev environment on an Ubuntu 14.04
> 64-bit desktop from the AWS marketplace running on EC2.
> 
> It would seem it's an obvious YES answer, but given the time investment
> I've been making I just wanted to be absolutely sure before continuing.
> 
> I thought it possible that maybe some errors, depending on their nature,
> can be overlooked, and that other developers may be doing that in practice.
> And hence perhaps I should as well to save time. Yes or No??
> 
> Thank you,
> 
> Martin


-
To unsubscribe, e-mail: common-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-dev-h...@hadoop.apache.org



Re: Yes/No newbie question on contributing

2016-07-26 Thread Sean Busbey
The current HowToContribute guide expressly tells folks that they
should ensure all the tests run and pass before and after their
change.

Sounds like we're due for an update if the expectation is now that
folks should be using -DskipTests and runs on particular modules.
Maybe we could instruct folks on running the same checks we'll do in
the automated precommit builds?

On Tue, Jul 26, 2016 at 1:47 PM, Vinod Kumar Vavilapalli
 wrote:
> The short answer is that it is expected to pass without any errors.
>
> On branch-2.x, that command passes cleanly without any errors though it takes 
> north of 10 minutes. Note that I run it with -DskipTests - you don’t want to 
> wait for all the unit tests to run, that’ll take too much time. I expect 
> trunk to be the same too.
>
> Which branch are you running this against? What errors are you seeing? If it 
> is unit-tests you are talking about, you can instead run with skipTests, run 
> only specific tests or all tests in the module you are touching, make sure 
> they pass and then let Jenkins infrastructure run the remaining tests when 
> you submit the patch.
>
> +Vinod
>
>> On Jul 26, 2016, at 11:41 AM, Martin Rosse  wrote:
>>
>> Hi,
>>
>> In the How To Contribute doc, it says:
>>
>> "Try getting the project to build and test locally before writing code"
>>
>> So, just to be 100% certain before I keep troubleshooting things, this
>> means I should be able to run
>>
>> mvn clean install -Pdist -Dtar
>>
>> without getting any failures or errors at all...none...zero, right?
>>
>> I am surprised at how long this is taking as errors keep cropping up.
>> Should I just expect it to really take many hours (already at 10+) to work
>> through these issues? I am setting up a dev environment on an Ubuntu 14.04
>> 64-bit desktop from the AWS marketplace running on EC2.
>>
>> It would seem it's an obvious YES answer, but given the time investment
>> I've been making I just wanted to be absolutely sure before continuing.
>>
>> I thought it possible that maybe some errors, depending on their nature,
>> can be overlooked, and that other developers may be doing that in practice.
>> And hence perhaps I should as well to save time. Yes or No??
>>
>> Thank you,
>>
>> Martin
>
>
> -
> To unsubscribe, e-mail: common-dev-unsubscr...@hadoop.apache.org
> For additional commands, e-mail: common-dev-h...@hadoop.apache.org
>



-- 
busbey

-
To unsubscribe, e-mail: common-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-dev-h...@hadoop.apache.org



Re: Yes/No newbie question on contributing

2016-07-26 Thread Martin Rosse
Thanks everyone...that helped. I'll go ahead and edit the Wiki to clarify
the expectation.

I got a successful build using:

~/code/hadoop$  mvn install -DskipTests

To respond to Vinod's questions:



I think the answer is trunk. I obtained the source code using:

git clone git://git.apache.org/hadoop.git

...and the pom.xml in my source says version 3.0.0-alpha1-SNAPSHOT, and I
haven't tried to do anything with branches yet.



You were right--without knowing any better I was running all the unit
testsso I came across several errors...one error that I was able to fix
was apparently due a newline in the etc/hosts file as I learned from
https://issues.apache.org/jira/browse/HADOOP-10888. After my fix, a
subsequent build passed that unit test. But then a subsequent build to that
build caused that same error again, even thought the newline was fixed.

Another error I got when running mvn install without -DskipTests is
described in https://issues.apache.org/jira/browse/HADOOP-12611. This is
the type of error I thought would be worthy of ignoring.

Thanks again for your time--much appreciated!

-Martin




On Tue, Jul 26, 2016 at 1:27 PM, Sean Busbey  wrote:

> The current HowToContribute guide expressly tells folks that they
> should ensure all the tests run and pass before and after their
> change.
>
> Sounds like we're due for an update if the expectation is now that
> folks should be using -DskipTests and runs on particular modules.
> Maybe we could instruct folks on running the same checks we'll do in
> the automated precommit builds?
>
> On Tue, Jul 26, 2016 at 1:47 PM, Vinod Kumar Vavilapalli
>  wrote:
> > The short answer is that it is expected to pass without any errors.
> >
> > On branch-2.x, that command passes cleanly without any errors though it
> takes north of 10 minutes. Note that I run it with -DskipTests - you don’t
> want to wait for all the unit tests to run, that’ll take too much time. I
> expect trunk to be the same too.
> >
> > Which branch are you running this against? What errors are you seeing?
> If it is unit-tests you are talking about, you can instead run with
> skipTests, run only specific tests or all tests in the module you are
> touching, make sure they pass and then let Jenkins infrastructure run the
> remaining tests when you submit the patch.
> >
> > +Vinod
> >
> >> On Jul 26, 2016, at 11:41 AM, Martin Rosse  wrote:
> >>
> >> Hi,
> >>
> >> In the How To Contribute doc, it says:
> >>
> >> "Try getting the project to build and test locally before writing
> code"
> >>
> >> So, just to be 100% certain before I keep troubleshooting things, this
> >> means I should be able to run
> >>
> >> mvn clean install -Pdist -Dtar
> >>
> >> without getting any failures or errors at all...none...zero, right?
> >>
> >> I am surprised at how long this is taking as errors keep cropping up.
> >> Should I just expect it to really take many hours (already at 10+) to
> work
> >> through these issues? I am setting up a dev environment on an Ubuntu
> 14.04
> >> 64-bit desktop from the AWS marketplace running on EC2.
> >>
> >> It would seem it's an obvious YES answer, but given the time investment
> >> I've been making I just wanted to be absolutely sure before continuing.
> >>
> >> I thought it possible that maybe some errors, depending on their nature,
> >> can be overlooked, and that other developers may be doing that in
> practice.
> >> And hence perhaps I should as well to save time. Yes or No??
> >>
> >> Thank you,
> >>
> >> Martin
> >
> >
> > -
> > To unsubscribe, e-mail: common-dev-unsubscr...@hadoop.apache.org
> > For additional commands, e-mail: common-dev-h...@hadoop.apache.org
> >
>
>
>
> --
> busbey
>


Re: Yes/No newbie question on contributing

2016-07-28 Thread Martin Rosse
I don't have permissions to edit the Wiki, but I've included a link below
to my proposed revisions to the How To Contribute page. As a reminder,
these changes are meant to make it clear that one does not need to run/pass
*all* project unit tests before starting to write code or submit a patch.
Knowing this as a newbie would have saved me a lot of time.

I'm not sure whether my edits cover the suggestion of instructing folks to
run the same checks as are done in the automated precommit builds...I don't
know what those checks are. And we had not concluded whether to instruct
folks as such or not...thoughts?

https://docs.google.com/document/d/1wvGFQ9SgELwCnPanmZ4FmN_uNLW5iyMkVIr3A4CNrTk/edit

Best,
Martin


On Tue, Jul 26, 2016 at 2:58 PM, Martin Rosse  wrote:

> Thanks everyone...that helped. I'll go ahead and edit the Wiki to clarify
> the expectation.
>
> I got a successful build using:
>
> ~/code/hadoop$  mvn install -DskipTests
>
> To respond to Vinod's questions:
>
> 
>
> I think the answer is trunk. I obtained the source code using:
>
> git clone git://git.apache.org/hadoop.git
>
> ...and the pom.xml in my source says version 3.0.0-alpha1-SNAPSHOT, and I
> haven't tried to do anything with branches yet.
>
> 
>
> You were right--without knowing any better I was running all the unit
> testsso I came across several errors...one error that I was able to fix
> was apparently due a newline in the etc/hosts file as I learned from
> https://issues.apache.org/jira/browse/HADOOP-10888. After my fix, a
> subsequent build passed that unit test. But then a subsequent build to that
> build caused that same error again, even thought the newline was fixed.
>
> Another error I got when running mvn install without -DskipTests is
> described in https://issues.apache.org/jira/browse/HADOOP-12611. This is
> the type of error I thought would be worthy of ignoring.
>
> Thanks again for your time--much appreciated!
>
> -Martin
>
>
>
>
> On Tue, Jul 26, 2016 at 1:27 PM, Sean Busbey  wrote:
>
>> The current HowToContribute guide expressly tells folks that they
>> should ensure all the tests run and pass before and after their
>> change.
>>
>> Sounds like we're due for an update if the expectation is now that
>> folks should be using -DskipTests and runs on particular modules.
>> Maybe we could instruct folks on running the same checks we'll do in
>> the automated precommit builds?
>>
>> On Tue, Jul 26, 2016 at 1:47 PM, Vinod Kumar Vavilapalli
>>  wrote:
>> > The short answer is that it is expected to pass without any errors.
>> >
>> > On branch-2.x, that command passes cleanly without any errors though it
>> takes north of 10 minutes. Note that I run it with -DskipTests - you don’t
>> want to wait for all the unit tests to run, that’ll take too much time. I
>> expect trunk to be the same too.
>> >
>> > Which branch are you running this against? What errors are you seeing?
>> If it is unit-tests you are talking about, you can instead run with
>> skipTests, run only specific tests or all tests in the module you are
>> touching, make sure they pass and then let Jenkins infrastructure run the
>> remaining tests when you submit the patch.
>> >
>> > +Vinod
>> >
>> >> On Jul 26, 2016, at 11:41 AM, Martin Rosse  wrote:
>> >>
>> >> Hi,
>> >>
>> >> In the How To Contribute doc, it says:
>> >>
>> >> "Try getting the project to build and test locally before writing
>> code"
>> >>
>> >> So, just to be 100% certain before I keep troubleshooting things, this
>> >> means I should be able to run
>> >>
>> >> mvn clean install -Pdist -Dtar
>> >>
>> >> without getting any failures or errors at all...none...zero, right?
>> >>
>> >> I am surprised at how long this is taking as errors keep cropping up.
>> >> Should I just expect it to really take many hours (already at 10+) to
>> work
>> >> through these issues? I am setting up a dev environment on an Ubuntu
>> 14.04
>> >> 64-bit desktop from the AWS marketplace running on EC2.
>> >>
>> >> It would seem it's an obvious YES answer, but given the time investment
>> >> I've been making I just wanted to be absolutely sure before continuing.
>> >>
>> >> I thought it possible that maybe some errors, depending on their
>> nature,
>> >> can be overlooked, and that other developers may be doing that in
>> practice.
>> >> And hence perhaps I should as well to save time. Yes or No??
>> >>
>> >> Thank you,
>> >>
>> >> Martin
>> >
>> >
>> > -
>> > To unsubscribe, e-mail: common-dev-unsubscr...@hadoop.apache.org
>> > For additional commands, e-mail: common-dev-h...@hadoop.apache.org
>> >
>>
>>
>>
>> --
>> busbey
>>
>
>


Re: Yes/No newbie question on contributing

2016-08-31 Thread Andrew Wang
Hello Martin,

Sorry for the slow followup here. If you'd like, someone can give you edit
permissions to the wiki so you can make changes yourself. In this case,
please provide your wiki username.

I also took a look at your document, and am having a hard time determining
the diff between it and what is on the wiki now. Your doc still says
"Please make sure that all unit tests succeed before constructing your
patch and that no new javac compiler warnings are introduced by your
patch", which I believe is inaccurate based on our discussion above.

Best,
Andrew

On Thu, Jul 28, 2016 at 4:59 PM, Martin Rosse  wrote:

> I don't have permissions to edit the Wiki, but I've included a link below
> to my proposed revisions to the How To Contribute page. As a reminder,
> these changes are meant to make it clear that one does not need to run/pass
> *all* project unit tests before starting to write code or submit a patch.
> Knowing this as a newbie would have saved me a lot of time.
>
> I'm not sure whether my edits cover the suggestion of instructing folks to
> run the same checks as are done in the automated precommit builds...I don't
> know what those checks are. And we had not concluded whether to instruct
> folks as such or not...thoughts?
>
> https://docs.google.com/document/d/1wvGFQ9SgELwCnPanmZ4FmN_
> uNLW5iyMkVIr3A4CNrTk/edit
>
> Best,
> Martin
>
>
> On Tue, Jul 26, 2016 at 2:58 PM, Martin Rosse  wrote:
>
> > Thanks everyone...that helped. I'll go ahead and edit the Wiki to clarify
> > the expectation.
> >
> > I got a successful build using:
> >
> > ~/code/hadoop$  mvn install -DskipTests
> >
> > To respond to Vinod's questions:
> >
> > 
> >
> > I think the answer is trunk. I obtained the source code using:
> >
> > git clone git://git.apache.org/hadoop.git
> >
> > ...and the pom.xml in my source says version 3.0.0-alpha1-SNAPSHOT, and I
> > haven't tried to do anything with branches yet.
> >
> > 
> >
> > You were right--without knowing any better I was running all the unit
> > testsso I came across several errors...one error that I was able to
> fix
> > was apparently due a newline in the etc/hosts file as I learned from
> > https://issues.apache.org/jira/browse/HADOOP-10888. After my fix, a
> > subsequent build passed that unit test. But then a subsequent build to
> that
> > build caused that same error again, even thought the newline was fixed.
> >
> > Another error I got when running mvn install without -DskipTests is
> > described in https://issues.apache.org/jira/browse/HADOOP-12611. This is
> > the type of error I thought would be worthy of ignoring.
> >
> > Thanks again for your time--much appreciated!
> >
> > -Martin
> >
> >
> >
> >
> > On Tue, Jul 26, 2016 at 1:27 PM, Sean Busbey 
> wrote:
> >
> >> The current HowToContribute guide expressly tells folks that they
> >> should ensure all the tests run and pass before and after their
> >> change.
> >>
> >> Sounds like we're due for an update if the expectation is now that
> >> folks should be using -DskipTests and runs on particular modules.
> >> Maybe we could instruct folks on running the same checks we'll do in
> >> the automated precommit builds?
> >>
> >> On Tue, Jul 26, 2016 at 1:47 PM, Vinod Kumar Vavilapalli
> >>  wrote:
> >> > The short answer is that it is expected to pass without any errors.
> >> >
> >> > On branch-2.x, that command passes cleanly without any errors though
> it
> >> takes north of 10 minutes. Note that I run it with -DskipTests - you
> don’t
> >> want to wait for all the unit tests to run, that’ll take too much time.
> I
> >> expect trunk to be the same too.
> >> >
> >> > Which branch are you running this against? What errors are you seeing?
> >> If it is unit-tests you are talking about, you can instead run with
> >> skipTests, run only specific tests or all tests in the module you are
> >> touching, make sure they pass and then let Jenkins infrastructure run
> the
> >> remaining tests when you submit the patch.
> >> >
> >> > +Vinod
> >> >
> >> >> On Jul 26, 2016, at 11:41 AM, Martin Rosse 
> wrote:
> >> >>
> >> >> Hi,
> >> >>
> >> >> In the How To Contribute doc, it says:
> >> >>
> >> >> "Try getting the project to build and test locally before writing
> >> code"
> >> >>
> >> >> So, just to be 100% certain before I keep troubleshooting things,
> this
> >> >> means I should be able to run
> >> >>
> >> >> mvn clean install -Pdist -Dtar
> >> >>
> >> >> without getting any failures or errors at all...none...zero, right?
> >> >>
> >> >> I am surprised at how long this is taking as errors keep cropping up.
> >> >> Should I just expect it to really take many hours (already at 10+) to
> >> work
> >> >> through these issues? I am setting up a dev environment on an Ubuntu
> >> 14.04
> >> >> 64-bit desktop from the AWS marketplace running on EC2.
> >> >>
> >> >> It would seem it's an obvious YES answer, but given the time
> investment
> >> >> I've been making I just wanted to be absolutely sure before
> continuing.
> >> >>
> >