Upgrading rspec?

[Luke Kanies]

Anyone have an issue with me upgrade the rspec version required by
Puppet's tests?

1.2.9 is out and seems to work fine, and someone *cough*Rein*cough*
wrote tests that required it somehow, so now I'm stuck with a conflict.

I don't see any reason not to use more recent ones...

--
The people who are regarded as moral luminaries are those who forego
ordinary pleasures themselves and find compensation in interfering
with the pleasures of others. -- Bertrand Russell
---------------------------------------------------------------------
Luke Kanies -|- http://reductivelabs.com -|- +1(615)594-8199

[br...@reductivelabs.com]

On Fri, Dec 4, 2009 at 3:25 PM, Luke Kanies <lu...@reductivelabs.com> wrote:

Anyone have an issue with me upgrade the rspec version required by
Puppet's tests?

1.2.9 is out and seems to work fine, and someone *cough*Rein*cough*
wrote tests that required it somehow, so now I'm stuck with a conflict.

I don't see any reason not to use more recent ones...

+1

I'd happily say goodbye to having to use the `spec _1.2.2_` hack when running files directly.

--
Bruce Williams
Developer @ Reductive Labs, http://reductivelabs.com

[Markus Roberts]

+1. I'm doing similar dodges to version mungle; it'd be nice to
standardize on the current version.

> --
>
> You received this message because you are subscribed to the Google Groups
> "Puppet Developers" group.
> To post to this group, send email to puppe...@googlegroups.com.
> To unsubscribe from this group, send email to
> puppet-dev+...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/puppet-dev?hl=en.
>

[Brice Figureau]

On 05/12/09 00:25, Luke Kanies wrote:
> Anyone have an issue with me upgrade the rspec version required by
> Puppet's tests?
>
> 1.2.9 is out and seems to work fine, and someone *cough*Rein*cough*
> wrote tests that required it somehow, so now I'm stuck with a conflict.
>
> I don't see any reason not to use more recent ones...

+1, I don't see why we're stuck on a precise version. I do prefer to
stick to the latest release and fix issues when they arise.

Can we have something in the rspec setup that warn if the used version
is known to not work?
--
Brice Figureau
My Blog: http://www.masterzen.fr/

[James Turnbull]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Brice Figureau wrote:
> +1, I don't see why we're stuck on a precise version. I do prefer to
> stick to the latest release and fix issues when they arise.
>
> Can we have something in the rspec setup that warn if the used version
> is known to not work?

+1. I can't even remember the exact issue that forced us to this
version - probably better to fix that issue than hold back RSpec.

Regards

James Turnbull

- --
Author of:
* Pro Linux System Administration (http://tinyurl.com/linuxadmin)
* Pulling Strings with Puppet (http://tinyurl.com/pupbook)
* Pro Nagios 2.0 (http://tinyurl.com/pronagios)
* Hardening Linux (http://tinyurl.com/hardeninglinux)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEVAwUBSxpBjCFa/lDkFHAyAQLnSwgAnvtGmtPhEaI87OyCh3zgE1qr1qQv5hHd
kx3GA8KOIqvQ4Bl98oaOpvdWRTO7f8tCw5tw5YErIhXMYyxnkpTwWj+is/tJ6rfI
hAlqaXsqW+oFTmSE0hR1/YYHewTzJTbEUox3dZwK6xEqwgOQLwbiPRWwcor6TfWc
ln31fC0vC04PLSmvTRvoDX5pR8M+1292Lhs2a9m351BaAMFaTbgxXeWrwc+tAexJ
LxiHqfVTOJMJE4RayJ5XbZpsAMP6dCCv4UPDNrW0ZB5+7LiK/kwV8q1HvJtczXMS
Egy5swrPdUSdGDPw4iIqCAx16hdipRPmhE03YKjPgE4GBszhJ+T6Qg==
=GTbO
-----END PGP SIGNATURE-----

[Luke Kanies]

On Dec 5, 2009, at 3:18 AM, James Turnbull wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Brice Figureau wrote:
>> +1, I don't see why we're stuck on a precise version. I do prefer to
>> stick to the latest release and fix issues when they arise.
>>
>> Can we have something in the rspec setup that warn if the used
>> version
>> is known to not work?
>
> +1. I can't even remember the exact issue that forced us to this
> version - probably better to fix that issue than hold back RSpec.

It wasn't a question of holding back but of forcing people to
upgrade. Rspec is almost as bad as rails in its API changes every
release, and we found that many people couldn't run our tests because
they hadn't upgraded.

Can we require at least a given version, using gems?

--
Those who speak most of progress measure it by quantity and not
by quality. --George Santayana
---------------------------------------------------------------------
Luke Kanies | http://reductivelabs.com | http://madstop.com

[Thomas Bellman]

Luke Kanies wrote:

> Anyone have an issue with me upgrade the rspec version required by
> Puppet's tests?
>
> 1.2.9 is out and seems to work fine, and someone *cough*Rein*cough*
> wrote tests that required it somehow, so now I'm stuck with a conflict.
>
> I don't see any reason not to use more recent ones...

Well, the latest one available in EPEL and in Fedora 12 is 1.2.7, so
requiring anything newer than that makes it difficult to run the tests
on those platforms.


/Bellman

[Markus Roberts]

AFAIK, the current version is available for all platforms (rspec is
pure ruby--nothing in it is platform specific) so there should be
nothing difficult about running the most current version on any
platform. All you should need to do is pull down the gem and go. As
this is something that you should be doing on a directly controlled
machine (i.e., you won't be needing to install rspec on the clients or
anything) I'm not seeing where this presents a difficulty.

-- Markus

[br...@reductivelabs.com]

From my perspective, for _development_ dependencies we shouldn't have to wait on system packages being updated -- even for dependencies with native code, as I think the use of RubyGems and build tools are a reasonable requirement for a development box.

Obviously this is not the case for production dependencies.

Cheers,

Bruce

[Luke Kanies]

I agree, and that's been our position on development dependencies so
far.

--
I have lost friends, some by death... others through sheer inability
to cross the street. -- Virginia Woolf

[Thomas Bellman]

Markus Roberts wrote:

>> Well, the latest one available in EPEL and in Fedora 12 is 1.2.7, so
>> requiring anything newer than that makes it difficult to run the tests
>> on those platforms.
>
> AFAIK, the current version is available for all platforms (rspec is
> pure ruby--nothing in it is platform specific) so there should be
> nothing difficult about running the most current version on any
> platform. All you should need to do is pull down the gem and go.

Anyone who seriously does systems administration should realize that
GEM (and CPAN for that matter) sucks badly. A secondary package
system that doesn't play together with the normal one.

> As this is something that you should be doing on a directly controlled
> machine (i.e., you won't be needing to install rspec on the clients or
> anything) I'm not seeing where this presents a difficulty.

It's not exactly encouraging occasional developers/patch writers to do
unit tests, since we will need to install a throw-away machine just to
run them.


/Bellman

[Luke Kanies]

On Dec 9, 2009, at 1:58 AM, Thomas Bellman wrote:

> Markus Roberts wrote:
>
>>> Well, the latest one available in EPEL and in Fedora 12 is 1.2.7, so
>>> requiring anything newer than that makes it difficult to run the
>>> tests
>>> on those platforms.
>>
>> AFAIK, the current version is available for all platforms (rspec is
>> pure ruby--nothing in it is platform specific) so there should be
>> nothing difficult about running the most current version on any
>> platform. All you should need to do is pull down the gem and go.
>
> Anyone who seriously does systems administration should realize that
> GEM (and CPAN for that matter) sucks badly. A secondary package
> system that doesn't play together with the normal one.

I definitely agree with that.

>> As this is something that you should be doing on a directly
>> controlled
>> machine (i.e., you won't be needing to install rspec on the clients
>> or
>> anything) I'm not seeing where this presents a difficulty.
>
> It's not exactly encouraging occasional developers/patch writers to do
> unit tests, since we will need to install a throw-away machine just to
> run them.

Hmm, that's a good point. I'd always figured your admin workstation
would be easy to have these gems installed on, but it sounds like
that's not the case.

What would we need to do to make this easier? Be more flexible on
rspec versions or something?

--
Men never do evil so completely and cheerfully as when they do it from a
religious conviction. --Blaise Pascal

[Brice Figureau]

On 09/12/09 18:53, Luke Kanies wrote:
> On Dec 9, 2009, at 1:58 AM, Thomas Bellman wrote:
>
>> Markus Roberts wrote:
>>> As this is something that you should be doing on a directly
>>> controlled
>>> machine (i.e., you won't be needing to install rspec on the clients
>>> or
>>> anything) I'm not seeing where this presents a difficulty.
>>
>> It's not exactly encouraging occasional developers/patch writers to do
>> unit tests, since we will need to install a throw-away machine just to
>> run them.
>
> Hmm, that's a good point. I'd always figured your admin workstation
> would be easy to have these gems installed on, but it sounds like
> that's not the case.
>
> What would we need to do to make this easier? Be more flexible on
> rspec versions or something?

Indeed, like allowing any version above a version number (ie >= 1.2.x)
would be a good start.
Note: we have the same issue with mocha.

[Bruce Williams]

On Dec 9, 2009, at 9:53 AM, Luke Kanies <lu...@madstop.com> wrote:

> On Dec 9, 2009, at 1:58 AM, Thomas Bellman wrote:
>
>> Markus Roberts wrote:
>>
>>>> Well, the latest one available in EPEL and in Fedora 12 is 1.2.7,
>>>> so
>>>> requiring anything newer than that makes it difficult to run the
>>>> tests
>>>> on those platforms.
>>>
>>> AFAIK, the current version is available for all platforms (rspec is
>>> pure ruby--nothing in it is platform specific) so there should be
>>> nothing difficult about running the most current version on any
>>> platform. All you should need to do is pull down the gem and go.
>>
>> Anyone who seriously does systems administration should realize that
>> GEM (and CPAN for that matter) sucks badly. A secondary package
>> system that doesn't play together with the normal one.
>
> I definitely agree with that.
>

I'm not sure how this relates to unit testing.

While I can imagine a full integration test suite, if platform
specific, should have better OS packaging system requirements, this
strikes me as a total hamstring to a development process in other
cases (ie, a systems administration constraint where one shouldn't
exist). Maybe the issue here is that the two should be separated.

>>> As this is something that you should be doing on a directly
>>> controlled
>>> machine (i.e., you won't be needing to install rspec on the clients
>>> or
>>> anything) I'm not seeing where this presents a difficulty.
>>
>> It's not exactly encouraging occasional developers/patch writers to
>> do
>> unit tests, since we will need to install a throw-away machine just
>> to
>> run them.
>
> Hmm, that's a good point. I'd always figured your admin workstation
> would be easy to have these gems installed on, but it sounds like
> that's not the case.
>
> What would we need to do to make this easier? Be more flexible on
> rspec versions or something?
>
> --
> Men never do evil so completely and cheerfully as when they do it
> from a
> religious conviction. --Blaise Pascal
> ---------------------------------------------------------------------
> Luke Kanies | http://reductivelabs.com | http://madstop.com
>

[Rein Henrichs]

I think an immediate and mutually satisfactory solution is to use a
">=" version specification that allows infrequent patchers using an
older version of Rspec to contribute easily (which we obviously want)
and allows frequent contributors to use their (most likely) up-to-date
Rspec version without undue frustration (which we obviously want).

Also, Rubygems and your normal packaging system certainly don't work
well together and, as I have suggested in the past[1], it is best to
not ask them to. Luckily, Rubygems can be installed in and used from a
user's home directory quite easily and with a minimum of
configuration[2], which should make it quite feasible for most people
to have their cake and eat it too. And by cake, I mean Rubygems.

[1] https://help.ubuntu.com/community/RubyOnRails
[2] http://docs.rubygems.org/read/chapter/3#page83

--
Rein Henrichs
http://reductivelabs.com

[Luke Kanies]

I agree; we should just switch to that for the versions. I wasn't
actually sure if it would work. :/

--
Hanlon's razor:
Never attribute to malice that which can adequately be explained by
stupidity.

[Michael Stahnke]

> Well, the latest one available in EPEL and in Fedora 12 is 1.2.7, so
> requiring anything newer than that makes it difficult to run the tests
> on those platforms.
>
>
>        /Bellman
>

Hey, I am the maintainer for rspec in EPEL. I just bumped it in F12
Friday. (should be in testing)
I will bump it in EPEL this week. Check epel-testing by the end of the week.


stahnma

[Michael Stahnke]

Built: http://koji.fedoraproject.org/koji/taskinfo?taskID=1872317
Should be in testing this week.