You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cxf.apache.org by Daniel Kulp <dk...@apache.org> on 2017/09/06 18:36:59 UTC
[DISCUSS] Status of 3.0.x...
Just wanted to start a quick discussion about 3.0.x. We’ve historically done work on the master and then supported two fixes branches. With 3.2.0 being voted on now, I’m not sure if we would branch the 3.2.x-fixes branch immediately or wait a bit (we have historically waited a bit). Are there major changes forthcoming that would warrant targeting 3.3 sooner? If not, I think I’d like to stick on master for at least one round of fixes just to stabilize the 3.2.0 as users migrate.
In any case, we’ve now supported the 3.0.x series for 2 years, 4 months. Would it make sense to announce that 3.0.15 will be the final scheduled release off that branch? Maybe wait one more iteration? We haven’t been back porting much anyway. I grabbed the “central download stats” from Nexus for the cxf-core module (pretty much any usage of CXF requires that) just to get an idea of what is used. The top 10 are:
3.0.3 18%
3.1.11 11%
3.1.6 10%
3.1.10 7%
3.1.12 6%
3.1.7 5%
3.1.8 5%
3.1.5 4%
3.1.4 4%
3.0.1 4%
As you can see, the only 3.0.x versions in the top 10 are 3.0.1 and 3.0.3. Since they are both ancient, I doubt anyone using them is planning on upgrading to new 3.0.x versions anyway. The good news is that the 3 latest 3.1.x versions are in the top 5. Shows that the 3.1.x folks are keeping up to date. The last two 3.0.x releases (.13 and .14) are numbers 24 and 23 on the list.
Anway, curious as to peoples thoughts….
--
Daniel Kulp
dkulp@apache.org - http://dankulp.com/blog
Talend Community Coder - http://coders.talend.com
Re: [DISCUSS] Status of 3.0.x...
Posted by Colm O hEigeartaigh <co...@apache.org>.
+1 as well to dropping 3.0.x, and keeping master on 3.2.x for a while.
Colm.
On Wed, Sep 6, 2017 at 11:18 PM, Andy McCright <j....@gmail.com>
wrote:
> +1 to keeping 3.2 on master until 3.2.1. I also agree with declaring end
> of support for 3.0.X (barring security vulnerabilities).
>
> Good discussion! Thanks!
>
> Andy
>
> On Wed, Sep 6, 2017 at 2:14 PM Andriy Redko <dr...@gmail.com> wrote:
>
> > I think that keeping 3.2.x on master would make sense, at least till
> > 3.2.1. As Dennis pointed out,
> > with Java 9 just a few weeks away we may branch off 3.2 later and work on
> > 3.3 (master) to make it good
> > Jigsaw citizen. Supporting only 3.1.x and dropping 3.0.x sounds
> > reasonable, +1 to that.
> >
> > DK> Just wanted to start a quick discussion about 3.0.x. We’ve
> > historically done work on the master and then
> > DK> supported two fixes branches. With 3.2.0 being voted on now, I’m
> not
> > sure if we would branch the 3.2.x-fixes
> > DK> branch immediately or wait a bit (we have historically waited a bit).
> > Are there major changes forthcoming that
> > DK> would warrant targeting 3.3 sooner? If not, I think I’d like to
> > stick on master for at least one round of fixes just to stabilize the
> 3.2.0
> > as users migrate.
> >
> > DK> In any case, we’ve now supported the 3.0.x series for 2 years, 4
> > months. Would it make sense to announce that
> > DK> 3.0.15 will be the final scheduled release off that branch? Maybe
> > wait one more iteration? We haven’t been back
> > DK> porting much anyway. I grabbed the “central download stats” from
> > Nexus for the cxf-core module (pretty much any
> > DK> usage of CXF requires that) just to get an idea of what is used.
> The
> > top 10 are:
> >
> > DK> 3.0.3 18%
> > DK> 3.1.11 11%
> > DK> 3.1.6 10%
> > DK> 3.1.10 7%
> > DK> 3.1.12 6%
> > DK> 3.1.7 5%
> > DK> 3.1.8 5%
> > DK> 3.1.5 4%
> > DK> 3.1.4 4%
> > DK> 3.0.1 4%
> >
> >
> > DK> As you can see, the only 3.0.x versions in the top 10 are 3.0.1 and
> > 3.0.3. Since they are both ancient, I doubt
> > DK> anyone using them is planning on upgrading to new 3.0.x versions
> > anyway. The good news is that the 3 latest 3.1.x
> > DK> versions are in the top 5. Shows that the 3.1.x folks are keeping up
> > to date. The last two 3.0.x releases (.13 and .14) are numbers 24 and
> 23
> > on the list.
> >
> > DK> Anway, curious as to peoples thoughts….
> >
> >
> >
> >
>
--
Colm O hEigeartaigh
Talend Community Coder
http://coders.talend.com
Re: [DISCUSS] Status of 3.0.x...
Posted by Andy McCright <j....@gmail.com>.
+1 to keeping 3.2 on master until 3.2.1. I also agree with declaring end
of support for 3.0.X (barring security vulnerabilities).
Good discussion! Thanks!
Andy
On Wed, Sep 6, 2017 at 2:14 PM Andriy Redko <dr...@gmail.com> wrote:
> I think that keeping 3.2.x on master would make sense, at least till
> 3.2.1. As Dennis pointed out,
> with Java 9 just a few weeks away we may branch off 3.2 later and work on
> 3.3 (master) to make it good
> Jigsaw citizen. Supporting only 3.1.x and dropping 3.0.x sounds
> reasonable, +1 to that.
>
> DK> Just wanted to start a quick discussion about 3.0.x. We’ve
> historically done work on the master and then
> DK> supported two fixes branches. With 3.2.0 being voted on now, I’m not
> sure if we would branch the 3.2.x-fixes
> DK> branch immediately or wait a bit (we have historically waited a bit).
> Are there major changes forthcoming that
> DK> would warrant targeting 3.3 sooner? If not, I think I’d like to
> stick on master for at least one round of fixes just to stabilize the 3.2.0
> as users migrate.
>
> DK> In any case, we’ve now supported the 3.0.x series for 2 years, 4
> months. Would it make sense to announce that
> DK> 3.0.15 will be the final scheduled release off that branch? Maybe
> wait one more iteration? We haven’t been back
> DK> porting much anyway. I grabbed the “central download stats” from
> Nexus for the cxf-core module (pretty much any
> DK> usage of CXF requires that) just to get an idea of what is used. The
> top 10 are:
>
> DK> 3.0.3 18%
> DK> 3.1.11 11%
> DK> 3.1.6 10%
> DK> 3.1.10 7%
> DK> 3.1.12 6%
> DK> 3.1.7 5%
> DK> 3.1.8 5%
> DK> 3.1.5 4%
> DK> 3.1.4 4%
> DK> 3.0.1 4%
>
>
> DK> As you can see, the only 3.0.x versions in the top 10 are 3.0.1 and
> 3.0.3. Since they are both ancient, I doubt
> DK> anyone using them is planning on upgrading to new 3.0.x versions
> anyway. The good news is that the 3 latest 3.1.x
> DK> versions are in the top 5. Shows that the 3.1.x folks are keeping up
> to date. The last two 3.0.x releases (.13 and .14) are numbers 24 and 23
> on the list.
>
> DK> Anway, curious as to peoples thoughts….
>
>
>
>
Re: [DISCUSS] Status of 3.0.x...
Posted by Sergey Beryozkin <sb...@gmail.com>.
Indeed, good plan...
Sergey
On 06/09/17 20:14, Andriy Redko wrote:
> I think that keeping 3.2.x on master would make sense, at least till 3.2.1. As Dennis pointed out,
> with Java 9 just a few weeks away we may branch off 3.2 later and work on 3.3 (master) to make it good
> Jigsaw citizen. Supporting only 3.1.x and dropping 3.0.x sounds reasonable, +1 to that.
>
> DK> Just wanted to start a quick discussion about 3.0.x. We’ve historically done work on the master and then
> DK> supported two fixes branches. With 3.2.0 being voted on now, I’m not sure if we would branch the 3.2.x-fixes
> DK> branch immediately or wait a bit (we have historically waited a bit). Are there major changes forthcoming that
> DK> would warrant targeting 3.3 sooner? If not, I think I’d like to stick on master for at least one round of fixes just to stabilize the 3.2.0 as users migrate.
>
> DK> In any case, we’ve now supported the 3.0.x series for 2 years, 4 months. Would it make sense to announce that
> DK> 3.0.15 will be the final scheduled release off that branch? Maybe wait one more iteration? We haven’t been back
> DK> porting much anyway. I grabbed the “central download stats” from Nexus for the cxf-core module (pretty much any
> DK> usage of CXF requires that) just to get an idea of what is used. The top 10 are:
>
> DK> 3.0.3 18%
> DK> 3.1.11 11%
> DK> 3.1.6 10%
> DK> 3.1.10 7%
> DK> 3.1.12 6%
> DK> 3.1.7 5%
> DK> 3.1.8 5%
> DK> 3.1.5 4%
> DK> 3.1.4 4%
> DK> 3.0.1 4%
>
>
> DK> As you can see, the only 3.0.x versions in the top 10 are 3.0.1 and 3.0.3. Since they are both ancient, I doubt
> DK> anyone using them is planning on upgrading to new 3.0.x versions anyway. The good news is that the 3 latest 3.1.x
> DK> versions are in the top 5. Shows that the 3.1.x folks are keeping up to date. The last two 3.0.x releases (.13 and .14) are numbers 24 and 23 on the list.
>
> DK> Anway, curious as to peoples thoughts….
>
>
>
Re: [DISCUSS] Status of 3.0.x...
Posted by Andriy Redko <dr...@gmail.com>.
I think that keeping 3.2.x on master would make sense, at least till 3.2.1. As Dennis pointed out,
with Java 9 just a few weeks away we may branch off 3.2 later and work on 3.3 (master) to make it good
Jigsaw citizen. Supporting only 3.1.x and dropping 3.0.x sounds reasonable, +1 to that.
DK> Just wanted to start a quick discussion about 3.0.x. We’ve historically done work on the master and then
DK> supported two fixes branches. With 3.2.0 being voted on now, I’m not sure if we would branch the 3.2.x-fixes
DK> branch immediately or wait a bit (we have historically waited a bit). Are there major changes forthcoming that
DK> would warrant targeting 3.3 sooner? If not, I think I’d like to stick on master for at least one round of fixes just to stabilize the 3.2.0 as users migrate.
DK> In any case, we’ve now supported the 3.0.x series for 2 years, 4 months. Would it make sense to announce that
DK> 3.0.15 will be the final scheduled release off that branch? Maybe wait one more iteration? We haven’t been back
DK> porting much anyway. I grabbed the “central download stats” from Nexus for the cxf-core module (pretty much any
DK> usage of CXF requires that) just to get an idea of what is used. The top 10 are:
DK> 3.0.3 18%
DK> 3.1.11 11%
DK> 3.1.6 10%
DK> 3.1.10 7%
DK> 3.1.12 6%
DK> 3.1.7 5%
DK> 3.1.8 5%
DK> 3.1.5 4%
DK> 3.1.4 4%
DK> 3.0.1 4%
DK> As you can see, the only 3.0.x versions in the top 10 are 3.0.1 and 3.0.3. Since they are both ancient, I doubt
DK> anyone using them is planning on upgrading to new 3.0.x versions anyway. The good news is that the 3 latest 3.1.x
DK> versions are in the top 5. Shows that the 3.1.x folks are keeping up to date. The last two 3.0.x releases (.13 and .14) are numbers 24 and 23 on the list.
DK> Anway, curious as to peoples thoughts….
Re: [DISCUSS] Status of 3.0.x...
Posted by Daniel Kulp <dk...@apache.org>.
> On Sep 6, 2017, at 3:03 PM, Dennis Kieselhorst <ma...@dekies.de> wrote:
> No 2.7.x version within the top 10? I think we should declare 3.0.15 as
> the last 3.0.x release. We can make an exception if security fixes are
> necessary. The projects I know either still use 2.x or already migrated
> to 3.x-latest.
I just checked cxf-core which didn’t exist in 2.7. (it was either cxf-api or cxf-core in 2.x). If I look at cxf-parent, 2.7.18 takes #7.
--
Daniel Kulp
dkulp@apache.org - http://dankulp.com/blog
Talend Community Coder - http://coders.talend.com
Re: [DISCUSS] Status of 3.0.x...
Posted by Dennis Kieselhorst <ma...@dekies.de>.
Hi!
> Just wanted to start a quick discussion about 3.0.x. We’ve historically done work on the master and then supported two fixes branches. With 3.2.0 being voted on now, I’m not sure if we would branch the 3.2.x-fixes branch immediately or wait a bit (we have historically waited a bit). Are there major changes forthcoming that would warrant targeting 3.3 sooner? If not, I think I’d like to stick on master for at least one round of fixes just to stabilize the 3.2.0 as users migrate.
Maybe for JDK 9? But currently I would wait and keep 3.2.x in master...
> In any case, we’ve now supported the 3.0.x series for 2 years, 4 months. Would it make sense to announce that 3.0.15 will be the final scheduled release off that branch? Maybe wait one more iteration? We haven’t been back porting much anyway. I grabbed the “central download stats” from Nexus for the cxf-core module (pretty much any usage of CXF requires that) just to get an idea of what is used. The top 10 are:
>
> 3.0.3 18%
> 3.1.11 11%
> 3.1.6 10%
> 3.1.10 7%
> 3.1.12 6%
> 3.1.7 5%
> 3.1.8 5%
> 3.1.5 4%
> 3.1.4 4%
> 3.0.1 4%
>
No 2.7.x version within the top 10? I think we should declare 3.0.15 as
the last 3.0.x release. We can make an exception if security fixes are
necessary. The projects I know either still use 2.x or already migrated
to 3.x-latest.
Cheers
Dennis