You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cloudstack.apache.org by David Nalley <da...@gnsa.us> on 2013/05/30 16:51:12 UTC
RPM building tests in jenkins
Hi folks:
I came across an interesting problem today, and think it's one that
deserves fixing.
I looked at our jenkins config and found that we build packages in a
manner that is different from how our documentation tells users to
build those packages. IMO while there may be more than one way to skin
the cat, we should at least be testing the manner we tell others to
use. (or perhaps changing the method we tell users to use to match the
tests we are doing)
Such disconnects between how our user base consumes ACS and how we
test it is bound to cause us problems.
--David
Re: RPM building tests in jenkins
Posted by David Nalley <da...@gnsa.us>.
On Mon, Jun 3, 2013 at 12:10 PM, Prasanna Santhanam <ts...@apache.org> wrote:
> On Fri, May 31, 2013 at 04:58:45AM -0400, David Nalley wrote:
>> On Fri, May 31, 2013 at 2:59 AM, Prasanna Santhanam <ts...@apache.org> wrote:
>> > On Thu, May 30, 2013 at 10:51:12AM -0400, David Nalley wrote:
>> >> Hi folks:
>> >>
>> >> I came across an interesting problem today, and think it's one that
>> >> deserves fixing.
>> >>
>> >> I looked at our jenkins config and found that we build packages in a
>> >> manner that is different from how our documentation tells users to
>> >> build those packages. IMO while there may be more than one way to skin
>> >> the cat, we should at least be testing the manner we tell others to
>> >> use. (or perhaps changing the method we tell users to use to match the
>> >> tests we are doing)
>> >>
>> >> Such disconnects between how our user base consumes ACS and how we
>> >> test it is bound to cause us problems.
>> >>
>> >
>> > What was different? The package.sh script in our repo was modified to
>> > take options but performs the same steps as does the package job on
>> > jenkins. We only write the full complete `rpmbuild` command on our
>> > jenkins job.
>> >
>>
>> How do you know they will always remain the same?
>> Being the same is not the point - we aren't testing the way we tell
>> users to do it. We wouldn't know if the way we tell users is broken or
>> not, because we aren't exercising that path. One of them needs to
>> change so that they are the same - even if they are the same in
>> effect.
>>
>
> Oh I fixed this btw. we're in sync with docs now.
>
> --
Awesome - thanks
--David
Re: RPM building tests in jenkins
Posted by Prasanna Santhanam <ts...@apache.org>.
On Fri, May 31, 2013 at 04:58:45AM -0400, David Nalley wrote:
> On Fri, May 31, 2013 at 2:59 AM, Prasanna Santhanam <ts...@apache.org> wrote:
> > On Thu, May 30, 2013 at 10:51:12AM -0400, David Nalley wrote:
> >> Hi folks:
> >>
> >> I came across an interesting problem today, and think it's one that
> >> deserves fixing.
> >>
> >> I looked at our jenkins config and found that we build packages in a
> >> manner that is different from how our documentation tells users to
> >> build those packages. IMO while there may be more than one way to skin
> >> the cat, we should at least be testing the manner we tell others to
> >> use. (or perhaps changing the method we tell users to use to match the
> >> tests we are doing)
> >>
> >> Such disconnects between how our user base consumes ACS and how we
> >> test it is bound to cause us problems.
> >>
> >
> > What was different? The package.sh script in our repo was modified to
> > take options but performs the same steps as does the package job on
> > jenkins. We only write the full complete `rpmbuild` command on our
> > jenkins job.
> >
>
> How do you know they will always remain the same?
> Being the same is not the point - we aren't testing the way we tell
> users to do it. We wouldn't know if the way we tell users is broken or
> not, because we aren't exercising that path. One of them needs to
> change so that they are the same - even if they are the same in
> effect.
>
Oh I fixed this btw. we're in sync with docs now.
--
Prasanna.,
------------------------
Powered by BigRock.com
Re: RPM building tests in jenkins
Posted by David Nalley <da...@gnsa.us>.
On Fri, May 31, 2013 at 2:59 AM, Prasanna Santhanam <ts...@apache.org> wrote:
> On Thu, May 30, 2013 at 10:51:12AM -0400, David Nalley wrote:
>> Hi folks:
>>
>> I came across an interesting problem today, and think it's one that
>> deserves fixing.
>>
>> I looked at our jenkins config and found that we build packages in a
>> manner that is different from how our documentation tells users to
>> build those packages. IMO while there may be more than one way to skin
>> the cat, we should at least be testing the manner we tell others to
>> use. (or perhaps changing the method we tell users to use to match the
>> tests we are doing)
>>
>> Such disconnects between how our user base consumes ACS and how we
>> test it is bound to cause us problems.
>>
>
> What was different? The package.sh script in our repo was modified to
> take options but performs the same steps as does the package job on
> jenkins. We only write the full complete `rpmbuild` command on our
> jenkins job.
>
How do you know they will always remain the same?
Being the same is not the point - we aren't testing the way we tell
users to do it. We wouldn't know if the way we tell users is broken or
not, because we aren't exercising that path. One of them needs to
change so that they are the same - even if they are the same in
effect.
--David
Re: RPM building tests in jenkins
Posted by Prasanna Santhanam <ts...@apache.org>.
On Thu, May 30, 2013 at 10:51:12AM -0400, David Nalley wrote:
> Hi folks:
>
> I came across an interesting problem today, and think it's one that
> deserves fixing.
>
> I looked at our jenkins config and found that we build packages in a
> manner that is different from how our documentation tells users to
> build those packages. IMO while there may be more than one way to skin
> the cat, we should at least be testing the manner we tell others to
> use. (or perhaps changing the method we tell users to use to match the
> tests we are doing)
>
> Such disconnects between how our user base consumes ACS and how we
> test it is bound to cause us problems.
>
What was different? The package.sh script in our repo was modified to
take options but performs the same steps as does the package job on
jenkins. We only write the full complete `rpmbuild` command on our
jenkins job.
--
Prasanna.,
------------------------
Powered by BigRock.com