Mailing-List: contact common-dev-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: common-dev@hadoop.apache.org
Received-SPF: pass (athena.apache.org: message received from 54.191.145.13
 which is an MX secondary for common-dev@hadoop.apache.org)
From: Brahma Reddy Battula <brahmareddy.battula@huawei.com>
To: "common-dev@hadoop.apache.org" <common-dev@hadoop.apache.org>
Subject: RE: IMPORTANT: testing patches for branches
Thread-Topic: IMPORTANT: testing patches for branches
Thread-Index: AQHQfTN0Yy4TLvBUP0KxCWv6nISa3p1ZGJoAgADWOICAAGtHgIAAFDcAgAD34L4=
Date: Fri, 24 Apr 2015 02:09:36 +0000
Message-ID: 
 <8AD4EE147886274A8B495D6AF407DF698124B6CE@szxeml510-mbx.china.huawei.com>
References: <ACEFED78-5640-4439-8F1C-26027214F7D7@altiscale.com>
	<8D5F7E3237B3ED47B84CF187BB17B66611B352B0@SHSMSX103.ccr.corp.intel.com>
	<8999F4C9-A56A-4CF3-AC46-053EA407625B@altiscale.com>
	<47887D7E-5782-4C31-A458-010DDD6229B7@hortonworks.com>,<CAFqyyL=Gc1wD467Zb1j-tyuAb6n32g7mPbqmE0+4hXbXmZQVBQ@mail.gmail.com>
In-Reply-To: 
 <CAFqyyL=Gc1wD467Zb1j-tyuAb6n32g7mPbqmE0+4hXbXmZQVBQ@mail.gmail.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

=0A=
Overall approach and summary is very good... well done =0A=
=0A=
Apart from steve pointed which I had also noticed, following one also i had=
 seen where I did not seen any clue..=0A=
=0A=
The patch artifact directory on has been removed!=0A=
This is a fatal error for test-patch.sh. Aborting.=0A=
Jenkins (node H8) information at https://builds.apache.org/job/PreCommit-YA=
RN-Build/7477/ may provide some hints.  ( did I miss here something )=0A=
=0A=
Atlast one suggestion is : can we remove whitespace check..?=0A=
=0A=
Thanks & Regards=0A=
Brahma Reddy Battula=0A=
=0A=
________________________________________=0A=
From: Sidharta Seethana [sidharta.apache@gmail.com]=0A=
Sent: Friday, April 24, 2015 12:27 AM=0A=
To: common-dev@hadoop.apache.org=0A=
Subject: Re: IMPORTANT: testing patches for branches=0A=
=0A=
About (3.) , a lot of the check style rules seem to be arcane/unnecessary.=
=0A=
Please see : https://issues.apache.org/jira/browse/HADOOP-11869=0A=
=0A=
On Thu, Apr 23, 2015 at 10:45 AM, Steve Loughran <stevel@hortonworks.com>=
=0A=
wrote:=0A=
=0A=
>=0A=
> 1. I really like the new patch process, especially the at-a-glance summar=
y=0A=
> 2. I think being -1 on whitespace is overkill; Its just part of my " git=
=0A=
> apply -p 0 -3 --verbose --whitespace=3Dfix  " action. Accordingly, I won'=
t=0A=
> reject patches on whitespace alone.=0A=
> 3. if checkstyle is complaining, how to track it down? As example, I don'=
t=0A=
> see much from:=0A=
>=0A=
> https://builds.apache.org/job/PreCommit-HADOOP-Build/6167/artifact/patchp=
rocess/checkstyle-result-diff.txt=0A=
>=0A=
>=0A=
> > On 23 Apr 2015, at 12:21, Allen Wittenauer <aw@altiscale.com> wrote:=0A=
> >=0A=
> >=0A=
> >=0A=
> >=0A=
> >=0A=
> > On Apr 22, 2015, at 11:34 PM, Zheng, Kai <kai.zheng@intel.com> wrote:=
=0A=
> >=0A=
> >> Hi Allen,=0A=
> >>=0A=
> >> This sounds great.=0A=
> >>=0A=
> >>>> Naming a patch foo-HDFS-7285.00.patch should get tested on the=0A=
> HDFS-7285 branch.=0A=
> >> Does it happen locally in developer's machine when running=0A=
> test-patch.sh, or also mean something in Hadoop Jenkins building when a=
=0A=
> JIRA becoming patch available? Thanks.=0A=
> >=0A=
> >=0A=
> >       Both, now that a fix has been committed last night (there was a=
=0A=
> bug in the Jenkins handling).=0A=
> >=0A=
> >       Given a patch name or URL, Jenkins and even running locally will=
=0A=
> try a few different methods to figure out which branch to use  out.  Note=
=0A=
> that a branch name of =91gitX=92 where X is a valid git reference also wo=
rks to=0A=
> force a patch to start at a particular commit.=0A=
> >=0A=
> >       For local use, you=92ll want to use a =91spare=92 copy of the sou=
rce=0A=
> tree via the =97basedir option and use the =97resetrepo flag.  That will =
enable=0A=
> Jenkins-like behavior and gives it permission to make modifications and=
=0A=
> effectively nuke any changes in the source tree you point it at.=0A=
> (Basically the opposite of the =97dirty-workspace flag).  If you want to=
=0A=
> force a branch (for whatever reason, including where the branch can=92t b=
e=0A=
> figured out), you can use the =97branch option.=0A=
> >=0A=
> >       If you don=92t use =97resetrepo, test-patch.sh will warn that it=
=0A=
> thinks the wrong branch is being used but will push on anyway.=0A=
> >=0A=
> >       In any case, the result of what it thinks the branch is/should be=
=0A=
> will be in the summary output at the bottom along with the git ref that i=
t=0A=
> specifically used for the test.=0A=
>=0A=
>=0A=