You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openwebbeans.apache.org by Gurkan Erdogdu <gu...@yahoo.com> on 2009/02/02 21:41:01 UTC

[CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Hi

The M1 release [VOTE] is canceled due to the some shortcomings. I will start a new VOTE after re-release process.

Thanks;

/Gurkan



      

Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Posted by Matthias Wessendorf <ma...@apache.org>.
On Tue, Feb 3, 2009 at 7:59 AM, Gurkan Erdogdu <gu...@yahoo.com> wrote:
> +1
> I will re-start release process today. I will run RAT before deploying to staging-repo.

you can integrate it on the maven build

>
> Mark,
> In the samples/guess/pom.xml, there is a myfaces JSF dependencies as *compile* scope. Is there any reason why its scope has not a *provided*? When the scope is *compile*, all dependencies are injected into WEB-INF/lib. I think that we mark the Java EE related libraries as *provided*.
>
> WDYT?
>
>
>
>
> ________________________________
> From: Mark Struberg <st...@yahoo.de>
> To: openwebbeans-dev@incubator.apache.org
> Sent: Monday, February 2, 2009 11:39:50 PM
> Subject: Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1
>
> but the release-tag will be on that branch in this case, isn't? So deleting it afterwards is maybe not such a good idea. Imho all releases must have a tag in the repo which exactly matches the released artifacts. But maybe I'm too fussy.
>
> Anyway, it's probably better I let Gurkan do the release and only make the changes in my local GIT archive.
>
> LieGrue,
> strub
>
>
> --- Kevan Miller <ke...@gmail.com> schrieb am Mo, 2.2.2009:
>
>> Von: Kevan Miller <ke...@gmail.com>
>> Betreff: Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1
>> An: openwebbeans-dev@incubator.apache.org
>> Datum: Montag, 2. Februar 2009, 22:32
>> On Feb 2, 2009, at 4:08 PM, Mark Struberg wrote:
>>
>> > Gurkan, Kevan, should I wait with the SPI
>> implementation as long as you do the release?
>>
>> I'd wait.
>>
>> One way you can handle this type of conflict, is to create
>> a release branch (e.g. svn copy trunk branches/1.0.0-M1) and
>> finalize the release. This allows new development to
>> continue, even if release voting takes a while...). Does
>> require merging of changes, if updates are made to the
>> release branch. Since you don't really plan on
>> supporting the M1 branch, you can delete it after the
>> release vote passes...
>>
>> --kevan
>
>
>



-- 
Matthias Wessendorf

blog: http://matthiaswessendorf.wordpress.com/
sessions: http://www.slideshare.net/mwessendorf
twitter: http://twitter.com/mwessendorf

Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Posted by Gurkan Erdogdu <gu...@yahoo.com>.
I think that, we do not depend on any web or Java EE server. There are two different dependencies for the sample guess project;

1* Java EE API dependencies (example; servlet-api, jsp-api, jsf-api, jta,jpa etc.)
2* Third party dependencies (jsf-facelet)

Moreover, we can not depend on any Java EE API implementation, like JSF-RI or Myfaces RI, just depend on Java EE API. 

IMHO, we can set the third party dependencies as *compile* and Java EE dependecies as *provided*. In the README.txt, I explain how to run samples.

WDYT?




________________________________
From: Mark Struberg <st...@yahoo.de>
To: openwebbeans-dev@incubator.apache.org
Sent: Tuesday, February 3, 2009 9:57:04 AM
Subject: Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

oh good point!

There is nowhere mentioned on which server the samples has to run (or I did not read it carefully enough).

Currently there are a lot of assumptions for the installations (basically all the dependencies the samples denote as 'provided'), so most people will fail to run the samples because they need to install those jars first.

On my notebook I did run via jetty, so I changed all API scopes to <scope>compile</scope> and all implementation jars to <scope>runtime</scope>.

So maybe I forgot to undo my change for MyFaces, but at least this is a good point for discussing the issue ;)

Imho samples have to run out of the box, without forcing the users to manually add jars, etc.

What I did sometimes is to set the scope via a ${webappscope} variable which is set to compile in the jetty profile and to runtime in a JBoss profile.

LieGrue,
strub


--- Gurkan Erdogdu <gu...@yahoo.com> schrieb am Di, 3.2.2009:

> Von: Gurkan Erdogdu <gu...@yahoo.com>
> Betreff: Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1
> An: openwebbeans-dev@incubator.apache.org
> Datum: Dienstag, 3. Februar 2009, 7:59
> +1 
> I will re-start release process today. I will run RAT
> before deploying to staging-repo.
> 
> Mark,
> In the samples/guess/pom.xml, there is a myfaces JSF
> dependencies as *compile* scope. Is there any reason why its
> scope has not a *provided*? When the scope is *compile*, all
> dependencies are injected into WEB-INF/lib. I think that we
> mark the Java EE related libraries as *provided*.
> 
> WDYT?
> 
> 
> 
> 
> ________________________________
> From: Mark Struberg <st...@yahoo.de>
> To: openwebbeans-dev@incubator.apache.org
> Sent: Monday, February 2, 2009 11:39:50 PM
> Subject: Re: AW: [CANCELED] Publish OpenWebBeans
> 1.0.0-incubating-M1
> 
> but the release-tag will be on that branch in this case,
> isn't? So deleting it afterwards is maybe not such a
> good idea. Imho all releases must have a tag in the repo
> which exactly matches the released artifacts. But maybe
> I'm too fussy.
> 
> Anyway, it's probably better I let Gurkan do the
> release and only make the changes in my local GIT archive.
> 
> LieGrue,
> strub
> 
> 
> --- Kevan Miller <ke...@gmail.com> schrieb am
> Mo, 2.2.2009:
> 
> > Von: Kevan Miller <ke...@gmail.com>
> > Betreff: Re: AW: [CANCELED] Publish OpenWebBeans
> 1.0.0-incubating-M1
> > An: openwebbeans-dev@incubator.apache.org
> > Datum: Montag, 2. Februar 2009, 22:32
> > On Feb 2, 2009, at 4:08 PM, Mark Struberg wrote:
> > 
> > > Gurkan, Kevan, should I wait with the SPI
> > implementation as long as you do the release?
> > 
> > I'd wait.
> > 
> > One way you can handle this type of conflict, is to
> create
> > a release branch (e.g. svn copy trunk
> branches/1.0.0-M1) and
> > finalize the release. This allows new development to
> > continue, even if release voting takes a while...).
> Does
> > require merging of changes, if updates are made to the
> > release branch. Since you don't really plan on
> > supporting the M1 branch, you can delete it after the
> > release vote passes...
> > 
> > --kevan


      

Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Posted by Mark Struberg <st...@yahoo.de>.
oh good point!

There is nowhere mentioned on which server the samples has to run (or I did not read it carefully enough).

Currently there are a lot of assumptions for the installations (basically all the dependencies the samples denote as 'provided'), so most people will fail to run the samples because they need to install those jars first.

On my notebook I did run via jetty, so I changed all API scopes to <scope>compile</scope> and all implementation jars to <scope>runtime</scope>.

So maybe I forgot to undo my change for MyFaces, but at least this is a good point for discussing the issue ;)

Imho samples have to run out of the box, without forcing the users to manually add jars, etc.

What I did sometimes is to set the scope via a ${webappscope} variable which is set to compile in the jetty profile and to runtime in a JBoss profile.

LieGrue,
strub


--- Gurkan Erdogdu <gu...@yahoo.com> schrieb am Di, 3.2.2009:

> Von: Gurkan Erdogdu <gu...@yahoo.com>
> Betreff: Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1
> An: openwebbeans-dev@incubator.apache.org
> Datum: Dienstag, 3. Februar 2009, 7:59
> +1 
> I will re-start release process today. I will run RAT
> before deploying to staging-repo.
> 
> Mark,
> In the samples/guess/pom.xml, there is a myfaces JSF
> dependencies as *compile* scope. Is there any reason why its
> scope has not a *provided*? When the scope is *compile*, all
> dependencies are injected into WEB-INF/lib. I think that we
> mark the Java EE related libraries as *provided*.
> 
> WDYT?
> 
> 
> 
> 
> ________________________________
> From: Mark Struberg <st...@yahoo.de>
> To: openwebbeans-dev@incubator.apache.org
> Sent: Monday, February 2, 2009 11:39:50 PM
> Subject: Re: AW: [CANCELED] Publish OpenWebBeans
> 1.0.0-incubating-M1
> 
> but the release-tag will be on that branch in this case,
> isn't? So deleting it afterwards is maybe not such a
> good idea. Imho all releases must have a tag in the repo
> which exactly matches the released artifacts. But maybe
> I'm too fussy.
> 
> Anyway, it's probably better I let Gurkan do the
> release and only make the changes in my local GIT archive.
> 
> LieGrue,
> strub
> 
> 
> --- Kevan Miller <ke...@gmail.com> schrieb am
> Mo, 2.2.2009:
> 
> > Von: Kevan Miller <ke...@gmail.com>
> > Betreff: Re: AW: [CANCELED] Publish OpenWebBeans
> 1.0.0-incubating-M1
> > An: openwebbeans-dev@incubator.apache.org
> > Datum: Montag, 2. Februar 2009, 22:32
> > On Feb 2, 2009, at 4:08 PM, Mark Struberg wrote:
> > 
> > > Gurkan, Kevan, should I wait with the SPI
> > implementation as long as you do the release?
> > 
> > I'd wait.
> > 
> > One way you can handle this type of conflict, is to
> create
> > a release branch (e.g. svn copy trunk
> branches/1.0.0-M1) and
> > finalize the release. This allows new development to
> > continue, even if release voting takes a while...).
> Does
> > require merging of changes, if updates are made to the
> > release branch. Since you don't really plan on
> > supporting the M1 branch, you can delete it after the
> > release vote passes...
> > 
> > --kevan


      

Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Posted by Gurkan Erdogdu <gu...@yahoo.com>.
+1 
I will re-start release process today. I will run RAT before deploying to staging-repo.

Mark,
In the samples/guess/pom.xml, there is a myfaces JSF dependencies as *compile* scope. Is there any reason why its scope has not a *provided*? When the scope is *compile*, all dependencies are injected into WEB-INF/lib. I think that we mark the Java EE related libraries as *provided*.

WDYT?




________________________________
From: Mark Struberg <st...@yahoo.de>
To: openwebbeans-dev@incubator.apache.org
Sent: Monday, February 2, 2009 11:39:50 PM
Subject: Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

but the release-tag will be on that branch in this case, isn't? So deleting it afterwards is maybe not such a good idea. Imho all releases must have a tag in the repo which exactly matches the released artifacts. But maybe I'm too fussy.

Anyway, it's probably better I let Gurkan do the release and only make the changes in my local GIT archive.

LieGrue,
strub


--- Kevan Miller <ke...@gmail.com> schrieb am Mo, 2.2.2009:

> Von: Kevan Miller <ke...@gmail.com>
> Betreff: Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1
> An: openwebbeans-dev@incubator.apache.org
> Datum: Montag, 2. Februar 2009, 22:32
> On Feb 2, 2009, at 4:08 PM, Mark Struberg wrote:
> 
> > Gurkan, Kevan, should I wait with the SPI
> implementation as long as you do the release?
> 
> I'd wait.
> 
> One way you can handle this type of conflict, is to create
> a release branch (e.g. svn copy trunk branches/1.0.0-M1) and
> finalize the release. This allows new development to
> continue, even if release voting takes a while...). Does
> require merging of changes, if updates are made to the
> release branch. Since you don't really plan on
> supporting the M1 branch, you can delete it after the
> release vote passes...
> 
> --kevan


      

Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Posted by Kevan Miller <ke...@gmail.com>.
On Feb 2, 2009, at 4:39 PM, Mark Struberg wrote:

> but the release-tag will be on that branch in this case, isn't? So  
> deleting it afterwards is maybe not such a good idea. Imho all  
> releases must have a tag in the repo which exactly matches the  
> released artifacts. But maybe I'm too fussy.

I probably left a few things unsaid... The release "branch" can be  
deleted, the release "tag" would be in tags/1.0.0-M1.

svn copy trunk/ branches/1.0.0-M1
perform release process (part of the release process will create a  
tags/1.0.0-M1 "tag"). The only difference is you're releasing from  
branches rather than trunk.

After vote passes:

svn delete branches/1.0.0-M1  (this is optional, but encouraged, since  
we don't really plan on supporting 1.0.0-M1).

--kevan


Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Posted by Mark Struberg <st...@yahoo.de>.
but the release-tag will be on that branch in this case, isn't? So deleting it afterwards is maybe not such a good idea. Imho all releases must have a tag in the repo which exactly matches the released artifacts. But maybe I'm too fussy.

Anyway, it's probably better I let Gurkan do the release and only make the changes in my local GIT archive.

LieGrue,
strub


--- Kevan Miller <ke...@gmail.com> schrieb am Mo, 2.2.2009:

> Von: Kevan Miller <ke...@gmail.com>
> Betreff: Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1
> An: openwebbeans-dev@incubator.apache.org
> Datum: Montag, 2. Februar 2009, 22:32
> On Feb 2, 2009, at 4:08 PM, Mark Struberg wrote:
> 
> > Gurkan, Kevan, should I wait with the SPI
> implementation as long as you do the release?
> 
> I'd wait.
> 
> One way you can handle this type of conflict, is to create
> a release branch (e.g. svn copy trunk branches/1.0.0-M1) and
> finalize the release. This allows new development to
> continue, even if release voting takes a while...). Does
> require merging of changes, if updates are made to the
> release branch. Since you don't really plan on
> supporting the M1 branch, you can delete it after the
> release vote passes...
> 
> --kevan


      

Re: AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Posted by Kevan Miller <ke...@gmail.com>.
On Feb 2, 2009, at 4:08 PM, Mark Struberg wrote:

> Gurkan, Kevan, should I wait with the SPI implementation as long as  
> you do the release?

I'd wait.

One way you can handle this type of conflict, is to create a release  
branch (e.g. svn copy trunk branches/1.0.0-M1) and finalize the  
release. This allows new development to continue, even if release  
voting takes a while...). Does require merging of changes, if updates  
are made to the release branch. Since you don't really plan on  
supporting the M1 branch, you can delete it after the release vote  
passes...

--kevan

AW: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1

Posted by Mark Struberg <st...@yahoo.de>.
Gurkan, Kevan, should I wait with the SPI implementation as long as you do the release?

LieGrue,
strub


--- Gurkan Erdogdu <gu...@yahoo.com> schrieb am Mo, 2.2.2009:

> Von: Gurkan Erdogdu <gu...@yahoo.com>
> Betreff: [CANCELED] Publish OpenWebBeans 1.0.0-incubating-M1
> An: openwebbeans-dev@incubator.apache.org
> Datum: Montag, 2. Februar 2009, 21:41
> Hi
> 
> The M1 release [VOTE] is canceled due to the some
> shortcomings. I will start a new VOTE after re-release
> process.
> 
> Thanks;
> 
> /Gurkan