Kernel Freeze: 7-Sep-2008

[Peter Knoop]

Hi everyone,

Ian's work on putting together the kernel has settled in nicely (and
he's hopefully enjoyed a nice, relaxing, well-deserved vacation!) so its
time to consider freezing the kernel in preparation for Sakai 2.6.  We
would like to propose that the kernel (K1) will be frozen on 7-Sep-2008.


Also, to keep things simple and manageable with the roll-out of K1, Ian
was initially the only committer in the kernel project.  As we move
forward, we need to expand this, as I'm sure Ian will be taking future
vacations :)  So, if you need commit in kernel, please send an email to
sakai-dev and svn-admins describing why you would like access.

If you have any questions, please let me know.

Thanks!

-peter
Sakai Project Coordinator
----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ray Davis]

I'll break the ice for integration test maintenance:

http://jira.sakaiproject.org/jira/browse/KNL-2

I also expect to work on two "user" tasks over the next few weeks. As
usual with complex changes, I plan to develop on a branch, and get
design review and customer testing before merging into trunk. (Given the
ever-present possibility of interruptions, they may push past the
proposed freeze date for 2.6, but those are the breaks.)

http://jira.sakaiproject.org/jira/browse/SAK-10868
http://jira.sakaiproject.org/jira/browse/SAK-14230

Finally, I'm responsible for a major change in the Component Manager for
2.6 and need to be ready to take the heat. :)

Thanks,
Ray

On 8/19/2008 5:52 AM, Knoop, Peter wrote:

> Hi everyone,
>
> Ian's work on putting together the kernel has settled in nicely (and
> he's hopefully enjoyed a nice, relaxing, well-deserved vacation!) so its
> time to consider freezing the kernel in preparation for Sakai 2.6. We
> would like to propose that the kernel (K1) will be frozen on 7-Sep-2008.
>
>
> Also, to keep things simple and manageable with the roll-out of K1, Ian
> was initially the only committer in the kernel project. As we move
> forward, we need to expand this, as I'm sure Ian will be taking future
> vacations :) So, if you need commit in kernel, please send an email to
> sakai-dev and svn-admins describing why you would like access.
>
> If you have any questions, please let me know.
>
> Thanks!
>
> -peter
> Sakai Project Coordinator

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Stephen Marquard]

Hi,

Two questions about the kernel:

1. Why do we need a new JIRA project for it? It's much easier to keep it in the existing project (and there are many issues still open that relate to components now in kernel)

2. Previously different people were responsible for commit access to projects now all inside kernel (e.g. content, authz, site, etc.). It's not clear how pre- and post- kernel commit access should work. Is there an assumption that anyone involved in any kernel project has commit access to the whole kernel now, or that the kernel commit list is being reconstituted from scratch, or that commit access will be managed at a lower level (e.g. kernel/authz, kernel/content) ?

Cheers
Stephen

>>> "Knoop, Peter" <knoop@...> 2008/08/19 02:52 PM >>>

Hi everyone,

Ian's work on putting together the kernel has settled in nicely (and
he's hopefully enjoyed a nice, relaxing, well-deserved vacation!) so its
time to consider freezing the kernel in preparation for Sakai 2.6.  We
would like to propose that the kernel (K1) will be frozen on 7-Sep-2008.


Also, to keep things simple and manageable with the roll-out of K1, Ian
was initially the only committer in the kernel project.  As we move
forward, we need to expand this, as I'm sure Ian will be taking future
vacations :)  So, if you need commit in kernel, please send an email to
sakai-dev and svn-admins describing why you would like access.

If you have any questions, please let me know.

Thanks!

-peter
Sakai Project Coordinator
----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ian Boston]

Ray,
"Finally, I'm responsible for a major change in the Component Manager  
for
>
> 2.6 and need to be ready to take the heat. :) "

I thought all of this was in 2.5.x  or was it post 2.5 ?


And,
Please can we discuss on list what is happening to integration  
testing in the kernel.
Large amounts of work in branches scare me because I don't want  
anyone to waste time without a shared design.
(since out jiras don't got to sakai-dev, in the jira isn'; visible  
enough)

eg
KNL-2 had left me with the impression that you wanted *me* to remove  
the test framework from kernel, allowing *you* to create something  
outside kernel, and then we would do something else, not dependent on  
the test framework inside kernel ?

We desperately need better coverage is < 1% inside the kernel at the  
moment. Frankly that bad :(.

Ian

On 19 Aug 2008, at 14:16, Ray Davis wrote:

> I'll break the ice for integration test maintenance:
>
> http://jira.sakaiproject.org/jira/browse/KNL-2
>
> I also expect to work on two "user" tasks over the next few weeks. As
> usual with complex changes, I plan to develop on a branch, and get
> design review and customer testing before merging into trunk.  
> (Given the
> ever-present possibility of interruptions, they may push past the
> proposed freeze date for 2.6, but those are the breaks.)
>
> http://jira.sakaiproject.org/jira/browse/SAK-10868
> http://jira.sakaiproject.org/jira/browse/SAK-14230
>
> Finally, I'm responsible for a major change in the Component  
> Manager for
> 2.6 and need to be ready to take the heat. :)
>
> Thanks,
> Ray
>
> On 8/19/2008 5:52 AM, Knoop, Peter wrote:
> > Hi everyone,
> >
> > Ian's work on putting together the kernel has settled in nicely (and
> > he's hopefully enjoyed a nice, relaxing, well-deserved vacation!)  
> so its
> > time to consider freezing the kernel in preparation for Sakai  
> 2.6. We
> > would like to propose that the kernel (K1) will be frozen on 7-
> Sep-2008.
> >
> >
> > Also, to keep things simple and manageable with the roll-out of  
> K1, Ian
> > was initially the only committer in the kernel project. As we move
> > forward, we need to expand this, as I'm sure Ian will be taking  
> future
> > vacations :) So, if you need commit in kernel, please send an  
> email to
> > sakai-dev and svn-admins describing why you would like access.
> >
> > If you have any questions, please let me know.
> >
> > Thanks!
> >
> > -peter
> > Sakai Project Coordinator
>
>
> This automatic notification message was sent by Sakai Collab  
> (https://collab.sakaiproject.org//portal) from the DG: Development  
> (a.k.a. sakai-dev) site.
> You can modify how you receive notifications at My Workspace >  
> Preferences.

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ian Boston]

It was suggested that there was a separate jira to focus work on the  
release which was at a separate pace from the sakai release, as the  
versions in jira wont make much sense.

Kernel commit will (IMHO) be managed at a single level "kernel" as  
there is not 1 component and about 4 jars for the whole of kernel.  
All the Impls are in the same source tree.

Ian

On 19 Aug 2008, at 14:56, Stephen Marquard wrote:

> Hi,
>
> Two questions about the kernel:
>
> 1. Why do we need a new JIRA project for it? It's much easier to  
> keep it in the existing project (and there are many issues still  
> open that relate to components now in kernel)
>
> 2. Previously different people were responsible for commit access  
> to projects now all inside kernel (e.g. content, authz, site,  
> etc.). It's not clear how pre- and post- kernel commit access  
> should work. Is there an assumption that anyone involved in any  
> kernel project has commit access to the whole kernel now, or that  
> the kernel commit list is being reconstituted from scratch, or that  
> commit access will be managed at a lower level (e.g. kernel/authz,  
> kernel/content) ?
>
> Cheers
> Stephen
>
> >>> "Knoop, Peter" <knoop@...> 2008/08/19 02:52 PM >>>
>
> Hi everyone,
>
> Ian's work on putting together the kernel has settled in nicely (and
> he's hopefully enjoyed a nice, relaxing, well-deserved vacation!)  
> so its
> time to consider freezing the kernel in preparation for Sakai 2.6. We
> would like to propose that the kernel (K1) will be frozen on 7-
> Sep-2008.
>
>
> Also, to keep things simple and manageable with the roll-out of K1,  
> Ian
> was initially the only committer in the kernel project. As we move
> forward, we need to expand this, as I'm sure Ian will be taking future
> vacations :) So, if you need commit in kernel, please send an email to
> sakai-dev and svn-admins describing why you would like access.
>
> If you have any questions, please let me know.
>
> Thanks!
>
> -peter
> Sakai Project Coordinator
> ----------------------
> This automatic notification message was sent by Sakai Collab  
> (https://collab.sakaiproject.org//portal) from the DG: Development  
> (a.k.a. sakai-dev) site.
> You can modify how you receive notifications at My Workspace >  
> Preferences.
>
>
> This automatic notification message was sent by Sakai Collab  
> (https://collab.sakaiproject.org//portal) from the DG: Development  
> (a.k.a. sakai-dev) site.
> You can modify how you receive notifications at My Workspace >  
> Preferences.

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[David Horwitz]

Thanks for the user alias changes :-)


David

Ray Davis wrote:

> I'll break the ice for integration test maintenance:
>
> http://jira.sakaiproject.org/jira/browse/KNL-2
>
> I also expect to work on two "user" tasks over the next few weeks. As
> usual with complex changes, I plan to develop on a branch, and get
> design review and customer testing before merging into trunk. (Given the
> ever-present possibility of interruptions, they may push past the
> proposed freeze date for 2.6, but those are the breaks.)
>
> http://jira.sakaiproject.org/jira/browse/SAK-10868
> http://jira.sakaiproject.org/jira/browse/SAK-14230
>
> Finally, I'm responsible for a major change in the Component Manager for
> 2.6 and need to be ready to take the heat. :)
>
> Thanks,
> Ray
>
> On 8/19/2008 5:52 AM, Knoop, Peter wrote:
> > Hi everyone,
> >
> > Ian's work on putting together the kernel has settled in nicely (and
> > he's hopefully enjoyed a nice, relaxing, well-deserved vacation!) so
> its
> > time to consider freezing the kernel in preparation for Sakai 2.6. We
> > would like to propose that the kernel (K1) will be frozen on
> 7-Sep-2008.
> >
> >
> > Also, to keep things simple and manageable with the roll-out of K1, Ian
> > was initially the only committer in the kernel project. As we move
> > forward, we need to expand this, as I'm sure Ian will be taking future
> > vacations :) So, if you need commit in kernel, please send an email to
> > sakai-dev and svn-admins describing why you would like access.
> >
> > If you have any questions, please let me know.
> >
> > Thanks!
> >
> > -peter
> > Sakai Project Coordinator
>
> ------------------------------------------------------------------------
>
> This automatic notification message was sent by Sakai Collab
> (https://collab.sakaiproject.org//portal) from the DG: Development
> (a.k.a. sakai-dev) site.
> You can modify how you receive notifications at My Workspace >
> Preferences.

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ray Davis]

On 8/19/2008 7:29 AM, Ian Boston wrote:
> Ray,
> "Finally, I'm responsible for a major change in the Component Manager for
>>
>> 2.6 and need to be ready to take the heat. :) "
>
> I thought all of this was in 2.5.x  or was it post 2.5 ?

No, it's always been aimed for 2.6. It just took a while to be merged
into trunk.

> And,
> Please can we discuss on list what is happening to integration testing
> in the kernel.
> Large amounts of work in branches scare me because I don't want anyone
> to waste time without a shared design.
> (since out jiras don't got to sakai-dev, in the jira isn'; visible enough)
>
> eg
> KNL-2 had left me with the impression that you wanted *me* to remove the
> test framework from kernel, allowing *you* to create something outside
> kernel, and then we would do something else, not dependent on the test
> framework inside kernel ?

The only reason I didn't assign that task to myself and immediately
start fixing the most critical problem -- the broken regression tests --
is that I didn't have privileges to do either of those things. As soon
as I found that out, I emailed Peter and the svn admins. Sorry for the
confusion -- I got pretty confused too.  :)

 > We desperately need better coverage is < 1% inside the kernel at the
 > moment. Frankly that bad :(.

I've been sending email out to sakai-dev, updating Confluence, and
adding JIRA tasks about progress on improving our programmatic tests and
how others can help. I expect there will be plenty more coming down the
wire. For example, once we've made some changes to the master POM, we
can come up with a way to run all long-running regression tests
(including integration tests) automatically as part of continuous builds
and the release process, which would prevent problems like KNL-2 from
happening.

Best,
Ray

>
>
> Ian
>
> On 19 Aug 2008, at 14:16, Ray Davis wrote:
>
>> I'll break the ice for integration test maintenance:
>>
>> http://jira.sakaiproject.org/jira/browse/KNL-2
>>
>> I also expect to work on two "user" tasks over the next few weeks. As
>> usual with complex changes, I plan to develop on a branch, and get
>> design review and customer testing before merging into trunk. (Given the
>> ever-present possibility of interruptions, they may push past the
>> proposed freeze date for 2.6, but those are the breaks.)
>>
>> http://jira.sakaiproject.org/jira/browse/SAK-10868
>> http://jira.sakaiproject.org/jira/browse/SAK-14230
>>
>> Finally, I'm responsible for a major change in the Component Manager for
>> 2.6 and need to be ready to take the heat. :)
>>
>> Thanks,
>> Ray
>>
>> On 8/19/2008 5:52 AM, Knoop, Peter wrote:
>> > Hi everyone,
>> >
>> > Ian's work on putting together the kernel has settled in nicely (and
>> > he's hopefully enjoyed a nice, relaxing, well-deserved vacation!) so
>> its
>> > time to consider freezing the kernel in preparation for Sakai 2.6. We
>> > would like to propose that the kernel (K1) will be frozen on
>> 7-Sep-2008.
>> >
>> >
>> > Also, to keep things simple and manageable with the roll-out of K1, Ian
>> > was initially the only committer in the kernel project. As we move
>> > forward, we need to expand this, as I'm sure Ian will be taking future
>> > vacations :) So, if you need commit in kernel, please send an email to
>> > sakai-dev and svn-admins describing why you would like access.
>> >
>> > If you have any questions, please let me know.
>> >
>> > Thanks!
>> >
>> > -peter
>> > Sakai Project Coordinator
>>
>>
>> This automatic notification message was sent by Sakai Collab
>> (https://collab.sakaiproject.org//portal) from the DG: Development
>> (a.k.a. sakai-dev) site.
>> You can modify how you receive notifications at My Workspace >
>> Preferences.
>
>

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ian Boston]

Ray,

Now that kernel is in one component and far simpler, it may be easier  
to fix inside kernel.

Please dont do anything just yet, I want to have a quick look.

Thanks
Ian




On 19 Aug 2008, at 15:54, Ray Davis wrote:

> On 8/19/2008 7:29 AM, Ian Boston wrote:
> > Ray,
> > "Finally, I'm responsible for a major change in the Component  
> Manager for
> >>
> >> 2.6 and need to be ready to take the heat. :) "
> >
> > I thought all of this was in 2.5.x or was it post 2.5 ?
>
> No, it's always been aimed for 2.6. It just took a while to be merged
> into trunk.
>
> > And,
> > Please can we discuss on list what is happening to integration  
> testing
> > in the kernel.
> > Large amounts of work in branches scare me because I don't want  
> anyone
> > to waste time without a shared design.
> > (since out jiras don't got to sakai-dev, in the jira isn';  
> visible enough)
> >
> > eg
> > KNL-2 had left me with the impression that you wanted *me* to  
> remove the
> > test framework from kernel, allowing *you* to create something  
> outside
> > kernel, and then we would do something else, not dependent on the  
> test
> > framework inside kernel ?
>
> The only reason I didn't assign that task to myself and immediately
> start fixing the most critical problem -- the broken regression  
> tests --
> is that I didn't have privileges to do either of those things. As soon
> as I found that out, I emailed Peter and the svn admins. Sorry for the
> confusion -- I got pretty confused too. :)
>
> > We desperately need better coverage is < 1% inside the kernel at the
> > moment. Frankly that bad :(.

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ray Davis]

Sure thing -- I have other work to keep me busy this week, and if the
broken tests become a blocker I can fix them on a temporary branch.

But since it's code that I've been actively developing, heavily rely on,
and know very well, please don't be reticent about working with me on
it, asking questions, etc. The K1 transition shouldn't leave you
personally responsible for managing every ongoing task related to K1
services.

I think you may have my IM info already, but drop me a line if you need it.

Best,
Ray

On 8/20/2008 3:58 AM, Ian Boston wrote:

> Ray,
>
> Now that kernel is in one component and far simpler, it may be easier to
> fix inside kernel.
>
> Please dont do anything just yet, I want to have a quick look.
>
> Thanks
> Ian
>
>
>
>
> On 19 Aug 2008, at 15:54, Ray Davis wrote:
>
>> On 8/19/2008 7:29 AM, Ian Boston wrote:
>> > Ray,
>> > "Finally, I'm responsible for a major change in the Component
>> Manager for
>> >>
>> >> 2.6 and need to be ready to take the heat. :) "
>> >
>> > I thought all of this was in 2.5.x or was it post 2.5 ?
>>
>> No, it's always been aimed for 2.6. It just took a while to be merged
>> into trunk.
>>
>> > And,
>> > Please can we discuss on list what is happening to integration testing
>> > in the kernel.
>> > Large amounts of work in branches scare me because I don't want anyone
>> > to waste time without a shared design.
>> > (since out jiras don't got to sakai-dev, in the jira isn'; visible
>> enough)
>> >
>> > eg
>> > KNL-2 had left me with the impression that you wanted *me* to remove
>> the
>> > test framework from kernel, allowing *you* to create something outside
>> > kernel, and then we would do something else, not dependent on the test
>> > framework inside kernel ?
>>
>> The only reason I didn't assign that task to myself and immediately
>> start fixing the most critical problem -- the broken regression tests --
>> is that I didn't have privileges to do either of those things. As soon
>> as I found that out, I emailed Peter and the svn admins. Sorry for the
>> confusion -- I got pretty confused too. :)
>>
>> > We desperately need better coverage is < 1% inside the kernel at the
>> > moment. Frankly that bad :(.
>
>

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ian Boston]

Ray,
I have re-written the test harness for kernel, and ported all the  
tests that were in kernel.

Here is a summary of the tests that are now running, although most  
are Running there are some failures, that I have a feeling are due to  
the tests being behind where they were in trunk.

So if you (or anyone on Sakai dev) recognized a test you wrote, a  
look at it, and a patch would be greatly appreciated.

BTW, these all ow run in eclipse OOTB, no config necessary. BUT the  
test harness uses a single classloader at the moment, which wont show  
up classloader errors.

Ian


-------------------------------------------------------
  T E S T S
-------------------------------------------------------
Running org.sakaiproject.user.impl.test.AuthenticationCacheTest
Tests run: 1, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 5.183  
sec <<< FAILURE!
Running org.sakaiproject.jcr.test.TestAll
Tests run: 1042, Failures: 0, Errors: 0, Skipped: 0, Time elapsed:  
27.224 sec
Running org.sakaiproject.component.test.DynamicConfigurationTest
Tests run: 2, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 5.004  
sec <<< FAILURE!
Running  
org.sakaiproject.content.impl.serialize.impl.test.ProfileSerializerTest
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 4.742  
sec
Running org.sakaiproject.email.impl.test.EmailServiceTest
Tests run: 6, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 4.921  
sec
Running  
org.sakaiproject.content.impl.serialize.impl.test.SaxSerializerTest
Tests run: 8, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.187  
sec
Running org.sakaiproject.user.impl.test.AuthenticatedUserProviderTest
Tests run: 4, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 4.996  
sec
Running org.sakaiproject.component.test.TestComponent
Tests run: 0, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.016  
sec
Running  
org.sakaiproject.user.impl.test.RequireLocalAccountLegacyAuthenticationT
est
Tests run: 2, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 4.906  
sec
Running org.sakaiproject.component.test.TestProvider1
Tests run: 0, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.016  
sec
Running org.sakaiproject.component.test.ConfigurationLoadingTest
Tests run: 2, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 6.117  
sec <<< FAILURE!
Running org.sakaiproject.component.test.TestProvider2
Tests run: 0, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.016  
sec
Running  
org.sakaiproject.content.impl.serialize.impl.test.Type1BaseContentCollec
tionSerializerTest
Tests run: 3, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.081  
sec
Running org.sakaiproject.user.impl.test.UserDirectoryServiceGetTest
Tests run: 12, Failures: 0, Errors: 0, Skipped: 0, Time elapsed:  
4.955 sec
Running  
org.sakaiproject.content.impl.serialize.impl.test.Type1BaseContentResour
ceSerializerTest
Tests run: 3, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.086  
sec
Running org.sakai.memory.impl.test.MemoryServiceTest
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.279  
sec
Running org.sakaiproject.authz.impl.test.AuthzIntegrationTest
Tests run: 2, Failures: 2, Errors: 0, Skipped: 0, Time elapsed: 4.891  
sec <<< FAILURE!
Running  
org.sakaiproject.content.impl.serialize.impl.test.GMTDateformatterTest
Tests run: 2, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.701  
sec

Results :

Failed tests:

Tests in error:

Tests run: 1091, Failures: 3, Errors: 2, Skipped: 0


On 20 Aug 2008, at 16:34, Ray Davis wrote:

> Sure thing -- I have other work to keep me busy this week, and if  
> the broken tests become a blocker I can fix them on a temporary  
> branch.
>
> But since it's code that I've been actively developing, heavily  
> rely on, and know very well, please don't be reticent about working  
> with me on it, asking questions, etc. The K1 transition shouldn't  
> leave you personally responsible for managing every ongoing task  
> related to K1 services.
>
> I think you may have my IM info already, but drop me a line if you  
> need it.
>
> Best,
> Ray

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ian Boston]

Ray,

I am not certain, but this might have been a very polite "can I have  
commit on /kernel"

Since you are already going through the process, this is early days  
and others may be uncertain about how to get commit on kernel, I  
would like to elaborate.

1. Commit on kernel allows a developer to work on the code directly.
2. Commit on kernel does not allow a developer design rights, that is  
done by discussion and specification.
3. Commit on kernel also implies some level of commitment on the  
developer to fix the code the developer has committed.

So how does a developer get commit on kernel ?

1. By contributing patches, and posts on list that the group of  
committers working on the code at the time feel indicate the the  
developer in question has
a) something to contribute to the aims of the kernel project
b) a collaborative attitude that sees the aims of the kernel project  
as more important than their own.

What other things might be needed in the Sakai context ?

1. The developer's manager giving some of their normal working time  
for the developer to dedicate to kernel work.
2. A vote by the current kernel committers off list. (all kernel  
committers are equal)

Consequently a request to commit indicates the developers wish to be  
considered.

---------------------------------------------------


As you know, you are being considered but while the group of  
committers is forming, this is all slightly artificial, but to ensure  
good practice, I am proposing that everyone experiences the same  
scrutiny regardless of past contributions. Obviously, being the first  
committer on kernel this leaves me in an odd position, although I  
feel the gestation of K1 gives me some justification to be the first  
kernel committer?

Ian


On 20 Aug 2008, at 16:34, Ray Davis wrote:

> The K1 transition shouldn't leave you personally responsible for  
> managing every ongoing task related to K1 services.

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Michael Korcuska]

I think Ian's proposal for getting commit to kernel is sound. We want  
to establish new/better/different standards for code contribution and  
the best way to do that is by starting from zero (ie past commit  
rights don't circumvent the process).

I was unclear about whether to apply this for the K1 work or to wait  
for K2, but I'm happier if everyone is willing to start now. I think  
we will need to clarify the procedures for K2 by trying to have some  
explicit coding standards we expect folks to follow (good work on this  
has already been done, which I would point to but I'm experiencing  
confluence issues right now) so it isn't simply a club, but rather a  
club with some rules that people follow and others know about.

Ian is in a unique position at this point, one of his own making, and  
I think we all would have been better served by a public conversation  
about what was happening and why *before* it happened. I (among others  
I'm sure) encouraged Ian to get the K1 work done before he left for  
vacation and we didn't explicitly discuss the issue of who would have  
commit. If you're upset at somebody how this transpired, it should be  
me.

That said, it is "the right thing to do" so we should just dust  
ourselves off and move forward, watching to make sure that any future  
process changes are discussed (or at least announced) in advance. To  
control quality and have consistent code it is most efficient/
effective to put someone as a gatekeeper. That somebody is Ian.

I'll encourage Ian to expand the group as quickly as possible, though,  
because its too much responsibility for a single person. And please  
also note that, whatever your seniority/skill level is, most of Ian's  
criteria deal with *how* the work is done in addition to the quality  
of the code itself. We're trying to establish a quality culture as  
well as quality code.

And, Ian, a big hearty thanks for all the effort you've put in to get  
us to this point. And thanks to everyone else for being willing to  
step back and "re-apply".

Michael

On Aug 26, 2008, at 10:12 AM, Ian Boston wrote:

> Ray,
>
> I am not certain, but this might have been a very polite "can I have
> commit on /kernel"
>
> Since you are already going through the process, this is early days
> and others may be uncertain about how to get commit on kernel, I
> would like to elaborate.
>
> 1. Commit on kernel allows a developer to work on the code directly.
> 2. Commit on kernel does not allow a developer design rights, that is
> done by discussion and specification.
> 3. Commit on kernel also implies some level of commitment on the
> developer to fix the code the developer has committed.
>
> So how does a developer get commit on kernel ?
>
> 1. By contributing patches, and posts on list that the group of
> committers working on the code at the time feel indicate the the
> developer in question has
> a) something to contribute to the aims of the kernel project
> b) a collaborative attitude that sees the aims of the kernel project
> as more important than their own.
>
> What other things might be needed in the Sakai context ?
>
> 1. The developer's manager giving some of their normal working time
> for the developer to dedicate to kernel work.
> 2. A vote by the current kernel committers off list. (all kernel
> committers are equal)
>
> Consequently a request to commit indicates the developers wish to be
> considered.
>
> ---------------------------------------------------
>
>
> As you know, you are being considered but while the group of
> committers is forming, this is all slightly artificial, but to ensure
> good practice, I am proposing that everyone experiences the same
> scrutiny regardless of past contributions. Obviously, being the first
> committer on kernel this leaves me in an odd position, although I
> feel the gestation of K1 gives me some justification to be the first
> kernel committer?
>
> Ian
>
>
> On 20 Aug 2008, at 16:34, Ray Davis wrote:
>
> > The K1 transition shouldn't leave you personally responsible for
> > managing every ongoing task related to K1 services.
>
>
> This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal 
> ) from the DG: Development (a.k.a. sakai-dev) site.
> You can modify how you receive notifications at My Workspace >  
> Preferences.

--
  Michael Korcuska
  Executive Director, Sakai Foundation
  mkorcuska@...
  mobile: [phone] // phone: [phone]
  skype: mkorcuska



----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Ray Davis]

OK, good to finally get word one way or the other. I'll submit patches
until further notice.

 > On Aug 26, 2008, at 10:12 AM, Ian Boston wrote:
 >
 >> Ray,
 >>
 >> I am not certain, but this might have been a very polite "can I have
 >> commit on /kernel"

My first request for commit access happened on August 12, and you were
CC-ed as project lead. My second was on August 19, in direct response to
Peter Knoop's message announcing that access requests were now
officially being considered, and sakai-dev was CC-ed.

Michael, I agree that the technical side of the K1 transition was
extremely well handled and that the policy change could have been better
communicated. The Apache model ensures that developers who are
responsible for a project and know it best get a chance to evaluate
potential new contributors while new contributors get a chance to learn
the codebase. This was instead an instance of a developer taking code
with which he wasn't familiar and relocating it in a broken state. Given
our existing 2.6 assignments and no word otherwise, I was as surprised
by Ian committing unreviewed changes to integration tests, User
Directory services, and the Component Manager as Ian would be by my
committing unreviewed changes to Content Hosting and memory cache services.

At any rate, thanks for the fuller explanation.

Ray

On 8/26/2008 5:53 AM, Michael Korcuska wrote:

> I think Ian's proposal for getting commit to kernel is sound. We want to
> establish new/better/different standards for code contribution and the
> best way to do that is by starting from zero (ie past commit rights
> don't circumvent the process).
>
> I was unclear about whether to apply this for the K1 work or to wait for
> K2, but I'm happier if everyone is willing to start now. I think we will
> need to clarify the procedures for K2 by trying to have some explicit
> coding standards we expect folks to follow (good work on this has
> already been done, which I would point to but I'm experiencing
> confluence issues right now) so it isn't simply a club, but rather a
> club with some rules that people follow and others know about.
>
> Ian is in a unique position at this point, one of his own making, and I
> think we all would have been better served by a public conversation
> about what was happening and why *before* it happened. I (among others
> I'm sure) encouraged Ian to get the K1 work done before he left for
> vacation and we didn't explicitly discuss the issue of who would have
> commit. If you're upset at somebody how this transpired, it should be me.
>
> That said, it is "the right thing to do" so we should just dust
> ourselves off and move forward, watching to make sure that any future
> process changes are discussed (or at least announced) in advance. To
> control quality and have consistent code it is most efficient/effective
> to put someone as a gatekeeper. That somebody is Ian.
>
> I'll encourage Ian to expand the group as quickly as possible, though,
> because its too much responsibility for a single person. And please also
> note that, whatever your seniority/skill level is, most of Ian's
> criteria deal with *how* the work is done in addition to the quality of
> the code itself. We're trying to establish a quality culture as well as
> quality code.
>
> And, Ian, a big hearty thanks for all the effort you've put in to get us
> to this point. And thanks to everyone else for being willing to step
> back and "re-apply".
>
> Michael
>
> On Aug 26, 2008, at 10:12 AM, Ian Boston wrote:
>
>> Ray,
>>
>> I am not certain, but this might have been a very polite "can I have
>> commit on /kernel"
>>
>> Since you are already going through the process, this is early days
>> and others may be uncertain about how to get commit on kernel, I
>> would like to elaborate.
>>
>> 1. Commit on kernel allows a developer to work on the code directly.
>> 2. Commit on kernel does not allow a developer design rights, that is
>> done by discussion and specification.
>> 3. Commit on kernel also implies some level of commitment on the
>> developer to fix the code the developer has committed.
>>
>> So how does a developer get commit on kernel ?
>>
>> 1. By contributing patches, and posts on list that the group of
>> committers working on the code at the time feel indicate the the
>> developer in question has
>> a) something to contribute to the aims of the kernel project
>> b) a collaborative attitude that sees the aims of the kernel project
>> as more important than their own.
>>
>> What other things might be needed in the Sakai context ?
>>
>> 1. The developer's manager giving some of their normal working time
>> for the developer to dedicate to kernel work.
>> 2. A vote by the current kernel committers off list. (all kernel
>> committers are equal)
>>
>> Consequently a request to commit indicates the developers wish to be
>> considered.
>>
>> ---------------------------------------------------
>>
>>
>> As you know, you are being considered but while the group of
>> committers is forming, this is all slightly artificial, but to ensure
>> good practice, I am proposing that everyone experiences the same
>> scrutiny regardless of past contributions. Obviously, being the first
>> committer on kernel this leaves me in an odd position, although I
>> feel the gestation of K1 gives me some justification to be the first
>> kernel committer?
>>
>> Ian
>>
>>
>> On 20 Aug 2008, at 16:34, Ray Davis wrote:
>>
>> > The K1 transition shouldn't leave you personally responsible for
>> > managing every ongoing task related to K1 services.
>>
>> ------------------------------------------------------------------------
>>
>> This automatic notification message was sent by Sakai Collab
>> (https://collab.sakaiproject.org//portal) from the DG: Development
>> (a.k.a. sakai-dev) site.
>> You can modify how you receive notifications at My Workspace >
>> Preferences.
>
> --
>  Michael Korcuska
>  Executive Director, Sakai Foundation
>  mkorcuska@... <mailto:mkorcuska@...>
>  mobile: [phone] // phone: [phone]
>  skype: mkorcuska
>
>

----------------------
This automatic notification message was sent by Sakai Collab (https://collab.sakaiproject.org//portal) from the DG: Development (a.k.a. sakai-dev) site.
You can modify how you receive notifications at My Workspace > Preferences.

[Michael Korcuska]

And I'll say that in re-reading my email I may have closed the door to  
discussion/dissent on the topic of how commit should be granted. Given  
the admiration of our group generally has for Apache (not to mention  
Apache's success), trying to model new commit privileges after Apache  
seems like a good idea. But we should definitely talk about whether  
this is the approach we want to take, especially for K2 and Sakai  
3.x.  I see applying this process to K1 as a dry-run for the K2/3.x  
commit process. So we're asking Ray (and others) to help establish  
processes for the future and set an example for others by going  
through a series of semi-formal steps. If others feel strongly that it  
is the wrong time to apply these steps or that it is the wrong process  
overall, please speak up (and hopefully suggest an alternative). I  
didn't mean to shut down discussion.

Ray is exactly right that one of the reasons for the Apache process is  
to let new developers get familiar with the code base and that reason  
doesn't apply in this case. But that's not the only reason for the  
process. The social and process aspects are, I suspect, just as  
important if not more so. There's also the idea of style consistency,  
etc. And lets remember Ian shouldn't be immune to these critiques  
either, which is another reason we need more than one committer--
someone to hold him accountable.


Thanks,

Michael

PS  Ray and Ian, I'd encourage you to connect synchronously about what  
happened with the changes surrounding K1. It's clear there is some  
tension there and I think getting whatever mutual frustrations you  
might have out in the open would be good. I know you both admire the  
other's technical abilities....perhaps a brief conversation about  
expectations and communication would help prevent this kind of thing  
in the future.

On Aug 26, 2008, at 6:08 PM, Ray Davis wrote:

> OK, good to finally get word one way or the other. I'll submit  
> patches until further notice.
>
&g