Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Ryan Tomayko]

On Sun, Jan 17, 2010 at 10:20 AM, Charles Oliver Nutter
<headius@headius.com> wrote:
> This came up because a friend of JRuby is working to prepare a Gentoo
> ebuild, and wants to get all related packages running green.

Speaking of, are there any friends of JRuby that are active on the
rack list? It'd be cool if someone could sign up for ensuring the test
suite is green before shipping new releases.

Thanks,
Ryan

Re: Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Charles Oliver Nutter]

If we get this green, I may advocate we set up a CI build for
JRuby+Rack somewhere. We've got a server (ci.jruby.org) where we could
put it, and it's obviously a pretty critical piece of infrastructure.

As for JRuby actives on the Rack lists...I'm not sure...

On Mon, Jan 18, 2010 at 12:42 AM, Ryan Tomayko <r@tomayko.com> wrote:
> On Sun, Jan 17, 2010 at 10:20 AM, Charles Oliver Nutter
> <headius@headius.com> wrote:
>> This came up because a friend of JRuby is working to prepare a Gentoo
>> ebuild, and wants to get all related packages running green.
>
> Speaking of, are there any friends of JRuby that are active on the
> rack list? It'd be cool if someone could sign up for ensuring the test
> suite is green before shipping new releases.
>
> Thanks,
> Ryan
>

Re: Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Joshua Peek]

+1 for any Rack CI. We have nothing right now.

On Mon, Jan 18, 2010 at 11:28 AM, Charles Oliver Nutter
<headius@headius.com> wrote:
> If we get this green, I may advocate we set up a CI build for
> JRuby+Rack somewhere. We've got a server (ci.jruby.org) where we could
> put it, and it's obviously a pretty critical piece of infrastructure.
>
> As for JRuby actives on the Rack lists...I'm not sure...
>
> On Mon, Jan 18, 2010 at 12:42 AM, Ryan Tomayko <r@tomayko.com> wrote:
>> On Sun, Jan 17, 2010 at 10:20 AM, Charles Oliver Nutter
>> <headius@headius.com> wrote:
>>> This came up because a friend of JRuby is working to prepare a Gentoo
>>> ebuild, and wants to get all related packages running green.
>>
>> Speaking of, are there any friends of JRuby that are active on the
>> rack list? It'd be cool if someone could sign up for ensuring the test
>> suite is green before shipping new releases.
>>
>> Thanks,
>> Ryan
>>
>



-- 
Joshua Peek

Re: Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Nick Sieger]

I'm on the Rack list as well, but (bad me) I haven't been diligent
enough to run Rack tests regularly.

I've got some CI set up for Rack with both Ruby 1.8.6 and JRuby trunk
at http://ci.jruby.org/view/Rack/, we'll see how they go.

/Nick

On Mon, Jan 18, 2010 at 11:28 AM, Charles Oliver Nutter
<headius@headius.com> wrote:
> If we get this green, I may advocate we set up a CI build for
> JRuby+Rack somewhere. We've got a server (ci.jruby.org) where we could
> put it, and it's obviously a pretty critical piece of infrastructure.
>
> As for JRuby actives on the Rack lists...I'm not sure...
>
> On Mon, Jan 18, 2010 at 12:42 AM, Ryan Tomayko <r@tomayko.com> wrote:
>> On Sun, Jan 17, 2010 at 10:20 AM, Charles Oliver Nutter
>> <headius@headius.com> wrote:
>>> This came up because a friend of JRuby is working to prepare a Gentoo
>>> ebuild, and wants to get all related packages running green.
>>
>> Speaking of, are there any friends of JRuby that are active on the
>> rack list? It'd be cool if someone could sign up for ensuring the test
>> suite is green before shipping new releases.
>>
>> Thanks,
>> Ryan
>>
>

Re: Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Charles Oliver Nutter]

Interesting failure out of the gate on Ruby 1.8.6...some sort of weird
log-processing issue?

On Mon, Jan 18, 2010 at 3:04 PM, Nick Sieger <nicksieger@gmail.com> wrote:
> I'm on the Rack list as well, but (bad me) I haven't been diligent
> enough to run Rack tests regularly.
>
> I've got some CI set up for Rack with both Ruby 1.8.6 and JRuby trunk
> at http://ci.jruby.org/view/Rack/, we'll see how they go.
>
> /Nick
>
> On Mon, Jan 18, 2010 at 11:28 AM, Charles Oliver Nutter
> <headius@headius.com> wrote:
>> If we get this green, I may advocate we set up a CI build for
>> JRuby+Rack somewhere. We've got a server (ci.jruby.org) where we could
>> put it, and it's obviously a pretty critical piece of infrastructure.
>>
>> As for JRuby actives on the Rack lists...I'm not sure...
>>
>> On Mon, Jan 18, 2010 at 12:42 AM, Ryan Tomayko <r@tomayko.com> wrote:
>>> On Sun, Jan 17, 2010 at 10:20 AM, Charles Oliver Nutter
>>> <headius@headius.com> wrote:
>>>> This came up because a friend of JRuby is working to prepare a Gentoo
>>>> ebuild, and wants to get all related packages running green.
>>>
>>> Speaking of, are there any friends of JRuby that are active on the
>>> rack list? It'd be cool if someone could sign up for ensuring the test
>>> suite is green before shipping new releases.
>>>
>>> Thanks,
>>> Ryan
>>>
>>
>

Re: Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Nick Sieger]

Yeah, anyone seen that before? For some reason the logger on the ci
box doesn't include the severity level in the output by default.

1) Failure:
test_spec {Rack::Logger} 001 [logs to rack.errors](Rack::Logger)
[./test/spec_rack_logger.rb:18]:
<"Program started\nNothing to do!\n"> expected to be =~
</INFO\ \-\-\ :\ Program\ started/>.

/Nick

On Mon, Jan 18, 2010 at 4:19 PM, Charles Oliver Nutter
<headius@headius.com> wrote:
> Interesting failure out of the gate on Ruby 1.8.6...some sort of weird
> log-processing issue?
>
> On Mon, Jan 18, 2010 at 3:04 PM, Nick Sieger <nicksieger@gmail.com> wrote:
>> I'm on the Rack list as well, but (bad me) I haven't been diligent
>> enough to run Rack tests regularly.
>>
>> I've got some CI set up for Rack with both Ruby 1.8.6 and JRuby trunk
>> at http://ci.jruby.org/view/Rack/, we'll see how they go.
>>
>> /Nick
>>
>> On Mon, Jan 18, 2010 at 11:28 AM, Charles Oliver Nutter
>> <headius@headius.com> wrote:
>>> If we get this green, I may advocate we set up a CI build for
>>> JRuby+Rack somewhere. We've got a server (ci.jruby.org) where we could
>>> put it, and it's obviously a pretty critical piece of infrastructure.
>>>
>>> As for JRuby actives on the Rack lists...I'm not sure...
>>>
>>> On Mon, Jan 18, 2010 at 12:42 AM, Ryan Tomayko <r@tomayko.com> wrote:
>>>> On Sun, Jan 17, 2010 at 10:20 AM, Charles Oliver Nutter
>>>> <headius@headius.com> wrote:
>>>>> This came up because a friend of JRuby is working to prepare a Gentoo
>>>>> ebuild, and wants to get all related packages running green.
>>>>
>>>> Speaking of, are there any friends of JRuby that are active on the
>>>> rack list? It'd be cool if someone could sign up for ensuring the test
>>>> suite is green before shipping new releases.
>>>>
>>>> Thanks,
>>>> Ryan
>>>>
>>>
>>
>

Re: Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Eric Wong]

Nick Sieger <nicksieger@gmail.com> wrote:
> Yeah, anyone seen that before? For some reason the logger on the ci
> box doesn't include the severity level in the output by default.
> 
> 1) Failure:
> test_spec {Rack::Logger} 001 [logs to rack.errors](Rack::Logger)
> [./test/spec_rack_logger.rb:18]:
> <"Program started\nNothing to do!\n"> expected to be =~
> </INFO\ \-\-\ :\ Program\ started/>.

Hi Nick,

I mentioned this in <20091231012252.GB7600@dcvr.yhbt.net>.

Running this test standalone works fine, so I suspect another test is
clobbering the default Logger::Formatter somewhere...

-- 
Eric Wong

Re: Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Nick Sieger]

On Tue, Jan 19, 2010 at 5:23 PM, Eric Wong <normalperson@yhbt.net> wrote:
> Nick Sieger <nicksieger@gmail.com> wrote:
>> Yeah, anyone seen that before? For some reason the logger on the ci
>> box doesn't include the severity level in the output by default.
>>
>> 1) Failure:
>> test_spec {Rack::Logger} 001 [logs to rack.errors](Rack::Logger)
>> [./test/spec_rack_logger.rb:18]:
>> <"Program started\nNothing to do!\n"> expected to be =~
>> </INFO\ \-\-\ :\ Program\ started/>.
>
> Hi Nick,
>
> I mentioned this in <20091231012252.GB7600@dcvr.yhbt.net>.
>
> Running this test standalone works fine, so I suspect another test is
> clobbering the default Logger::Formatter somewhere...

To see the failure, you need to install Camping and run the Camping
handler tests. The problem is Camping pulls in ActiveSupport, which
messes with Logger and overrides the default formatter.

The attached patch fixes the tests to make them insensitive to logger format.

/Nick

Re: Friends of Rack ∩ Friends of JRuby (was Re: Rack test spec_rack_runtime fails on JRuby due to timing)

[Nick Sieger]

[-- Attachment #1: Type: text/plain, Size: 1124 bytes --]

On Wed, Jan 20, 2010 at 12:08 PM, Nick Sieger <nicksieger@gmail.com> wrote:
> On Tue, Jan 19, 2010 at 5:23 PM, Eric Wong <normalperson@yhbt.net> wrote:
>> Nick Sieger <nicksieger@gmail.com> wrote:
>>> Yeah, anyone seen that before? For some reason the logger on the ci
>>> box doesn't include the severity level in the output by default.
>>>
>>> 1) Failure:
>>> test_spec {Rack::Logger} 001 [logs to rack.errors](Rack::Logger)
>>> [./test/spec_rack_logger.rb:18]:
>>> <"Program started\nNothing to do!\n"> expected to be =~
>>> </INFO\ \-\-\ :\ Program\ started/>.
>>
>> Hi Nick,
>>
>> I mentioned this in <20091231012252.GB7600@dcvr.yhbt.net>.
>>
>> Running this test standalone works fine, so I suspect another test is
>> clobbering the default Logger::Formatter somewhere...
>
> To see the failure, you need to install Camping and run the Camping
> handler tests. The problem is Camping pulls in ActiveSupport, which
> messes with Logger and overrides the default formatter.
>
> The attached patch fixes the tests to make them insensitive to logger format.

Stupid gmail ignored my attachment, here's another try.

/Nick

[-- Attachment #2: 0001-Make-logger-specs-insensitive-to-format-ActiveSup.patch --]
[-- Type: application/octet-stream, Size: 901 bytes --]

From f3bb696797f947c16b11f075f25365529a8f18fb Mon Sep 17 00:00:00 2001
From: Nick Sieger <nick@nicksieger.com>
Date: Wed, 20 Jan 2010 11:56:21 -0600
Subject: [PATCH] Make logger specs insensitive to format -- ActiveSupport (pulled in by Camping) messes with the logger formatter

---
 test/spec_rack_logger.rb |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/test/spec_rack_logger.rb b/test/spec_rack_logger.rb
index d55b9c7..2033006 100644
--- a/test/spec_rack_logger.rb
+++ b/test/spec_rack_logger.rb
@@ -15,7 +15,7 @@ context "Rack::Logger" do
 
     errors = StringIO.new
     Rack::Logger.new(app).call({'rack.errors' => errors})
-    errors.string.should.match "INFO -- : Program started"
-    errors.string.should.match "WARN -- : Nothing to do"
+    errors.string.should.match "Program started"
+    errors.string.should.match "Nothing to do"
   end
 end
-- 
1.5.6