All of lore.kernel.org
 help / color / mirror / Atom feed
* meta-java - recent commits break foreign (MY!) builds ...
@ 2015-08-24 16:48 Jens Rehsack
  2015-08-27 13:28 ` Otavio Salvador
  0 siblings, 1 reply; 7+ messages in thread
From: Jens Rehsack @ 2015-08-24 16:48 UTC (permalink / raw)
  To: OE-devel; +Cc: Tobias Theophile, Otavio Salvador, Mario Domenech Goulart

Hi,

recent commits (24b98ac3, 15ed9662) on meta-java broke my build, so I reverted them for my fork and decided to report:

15ed9662

This one is easy - and very likely similar easy to fix for patch author:

The layer-readme tells users:

Usage instructions
------------------

You should define at least the following variables in a distro include file or local.conf

PREFERRED_PROVIDER_virtual/java-initial-native = "cacao-initial-native"

PREFERRED_PROVIDER_virtual/java-native = "jamvm-native"
PREFERRED_PROVIDER_virtual/javac-native = "ecj-bootstrap-native"

With 15ed9662, jamvm requires ecj-bootstrap-native and ecj-bootstrap-native requires virtual/java-native which shall be jamvm-native.
Since I trust in you have a running version, there is very likely a way out by a nice and proper PREFERRED_PROVIDER_* setup making 15ed9662 succeed. 

24b98ac3

This is more complicated. I announced that in a small round already - and without this
commit I would hold back until I finished openjdk-8 cross-compiling and running on our
i.MX6 box with Framebuffer-XBMC and Java-Service for Smart-Home control.

See https://github.com/rehsack/meta-java/commits/master for an icedtea7-native and an
openjdk-8-native without any alsa or x11 dependency, compiling smoothly alongside a
gpu-viv-bin-mx6q (with a tiny patch removing hidden x11 dependency).

There is no serious need for x11 the dependency chain down from icedtea7-native.

Maybe the requirement as added was intended for target packages only, but hit accidentally
some native packages either.

Cheers
-- 
Jens Rehsack - rehsack@gmail.com



^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: meta-java - recent commits break foreign (MY!) builds ...
  2015-08-24 16:48 meta-java - recent commits break foreign (MY!) builds Jens Rehsack
@ 2015-08-27 13:28 ` Otavio Salvador
  2015-08-28  6:59   ` Richard Leitner
  0 siblings, 1 reply; 7+ messages in thread
From: Otavio Salvador @ 2015-08-27 13:28 UTC (permalink / raw)
  To: OpenEmbedded Devel List
  Cc: Tobias Theophile, Otavio Salvador, Mario Domenech Goulart

On Mon, Aug 24, 2015 at 1:48 PM, Jens Rehsack <rehsack@gmail.com> wrote:
\> recent commits (24b98ac3, 15ed9662) on meta-java broke my build, so
I reverted them for my fork and decided to report:
>
> 15ed9662
>
> This one is easy - and very likely similar easy to fix for patch author:
>
> The layer-readme tells users:
>
> Usage instructions
> ------------------
>
> You should define at least the following variables in a distro include file or local.conf
>
> PREFERRED_PROVIDER_virtual/java-initial-native = "cacao-initial-native"
>
> PREFERRED_PROVIDER_virtual/java-native = "jamvm-native"
> PREFERRED_PROVIDER_virtual/javac-native = "ecj-bootstrap-native"
>
> With 15ed9662, jamvm requires ecj-bootstrap-native and ecj-bootstrap-native requires virtual/java-native which shall be jamvm-native.
> Since I trust in you have a running version, there is very likely a way out by a nice and proper PREFERRED_PROVIDER_* setup making 15ed9662 succeed.

I ended reverting this one.

Can you help us to find a proper fix for this? We ended breaking this
as we were enabling the use of Oracle Java as builder and OpenJDK
ended broken.

> 24b98ac3
>
> This is more complicated. I announced that in a small round already - and without this
> commit I would hold back until I finished openjdk-8 cross-compiling and running on our
> i.MX6 box with Framebuffer-XBMC and Java-Service for Smart-Home control.

Yes, we need to fix it and to be honest I personally would kill
openjdk 7 once 8 is ready.

> See https://github.com/rehsack/meta-java/commits/master for an icedtea7-native and an
> openjdk-8-native without any alsa or x11 dependency, compiling smoothly alongside a
> gpu-viv-bin-mx6q (with a tiny patch removing hidden x11 dependency).

Can you prepare a patch for fsl-arm? we can review and fix it fast :)

> There is no serious need for x11 the dependency chain down from icedtea7-native.
>
> Maybe the requirement as added was intended for target packages only, but hit accidentally
> some native packages either.

You are right; we need _class-target on those. Can you fix it or want
me to cook a patch?

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: meta-java - recent commits break foreign (MY!) builds ...
  2015-08-27 13:28 ` Otavio Salvador
@ 2015-08-28  6:59   ` Richard Leitner
  2015-08-28 14:11     ` Otavio Salvador
  0 siblings, 1 reply; 7+ messages in thread
From: Richard Leitner @ 2015-08-28  6:59 UTC (permalink / raw)
  To: openembedded-devel; +Cc: Otavio Salvador

On 08/27/2015 03:28 PM, Otavio Salvador wrote:
> 
> Yes, we need to fix it and to be honest I personally would kill
> openjdk 7 once 8 is ready.
> 
I think we should keep (and UPDATE!) openjdk 6 & 7 as long as they get
support from RedHat and therefore IcedTea. There may be users relying on
the old java version for their long-term-supported products (so do we)!

IMHO we should start with openjdk 6: remove the old versions and add the
latest (1.13.8)!

What do you think? Any comments?

regards,
Richard L


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: meta-java - recent commits break foreign (MY!) builds ...
  2015-08-28  6:59   ` Richard Leitner
@ 2015-08-28 14:11     ` Otavio Salvador
  2015-08-31  7:33       ` Richard Leitner
  0 siblings, 1 reply; 7+ messages in thread
From: Otavio Salvador @ 2015-08-28 14:11 UTC (permalink / raw)
  To: OpenEmbedded Devel List; +Cc: Otavio Salvador

On Fri, Aug 28, 2015 at 3:59 AM, Richard Leitner
<richard.leitner@skidata.com> wrote:
> On 08/27/2015 03:28 PM, Otavio Salvador wrote:
>>
>> Yes, we need to fix it and to be honest I personally would kill
>> openjdk 7 once 8 is ready.
>>
> I think we should keep (and UPDATE!) openjdk 6 & 7 as long as they get
> support from RedHat and therefore IcedTea. There may be users relying on
> the old java version for their long-term-supported products (so do we)!
>
> IMHO we should start with openjdk 6: remove the old versions and add the
> latest (1.13.8)!
>
> What do you think? Any comments?

If you are willing to work on that, I have no problem.

The point is Java is complex and we have few people working on this so
I prefer one well maintained version than 3 oddly ones.

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: meta-java - recent commits break foreign (MY!) builds ...
  2015-08-28 14:11     ` Otavio Salvador
@ 2015-08-31  7:33       ` Richard Leitner
  2015-08-31 13:06         ` Otavio Salvador
  0 siblings, 1 reply; 7+ messages in thread
From: Richard Leitner @ 2015-08-31  7:33 UTC (permalink / raw)
  To: openembedded-devel; +Cc: Otavio Salvador

On 08/28/2015 04:11 PM, Otavio Salvador wrote:
> On Fri, Aug 28, 2015 at 3:59 AM, Richard Leitner
> <richard.leitner@skidata.com> wrote:
>> I think we should keep (and UPDATE!) openjdk 6 & 7 as long as they get
>> support from RedHat and therefore IcedTea. There may be users relying on
>> the old java version for their long-term-supported products (so do we)!
>>
>> IMHO we should start with openjdk 6: remove the old versions and add the
>> latest (1.13.8)!
>>
>> What do you think? Any comments?
> 
> If you are willing to work on that, I have no problem.

As I mentioned, I will update openjdk-7 as long as I need it. If it's
okay for you I will send patches for these updates to the ml. From a
current point of view it looks like I will provide updates until end of
2016 at least.

> 
> The point is Java is complex and we have few people working on this so
> I prefer one well maintained version than 3 oddly ones.

I understand that. But currently I have no need for openjdk-8. Therefore
my focus is on keeping openjdk-7 up to date.



^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: meta-java - recent commits break foreign (MY!) builds ...
  2015-08-31  7:33       ` Richard Leitner
@ 2015-08-31 13:06         ` Otavio Salvador
  2015-08-31 13:39           ` Richard Leitner
  0 siblings, 1 reply; 7+ messages in thread
From: Otavio Salvador @ 2015-08-31 13:06 UTC (permalink / raw)
  To: OpenEmbedded Devel List; +Cc: Otavio Salvador

On Mon, Aug 31, 2015 at 4:33 AM, Richard Leitner
<richard.leitner@skidata.com> wrote:
> On 08/28/2015 04:11 PM, Otavio Salvador wrote:
>> On Fri, Aug 28, 2015 at 3:59 AM, Richard Leitner
>> <richard.leitner@skidata.com> wrote:
>>> I think we should keep (and UPDATE!) openjdk 6 & 7 as long as they get
>>> support from RedHat and therefore IcedTea. There may be users relying on
>>> the old java version for their long-term-supported products (so do we)!
>>>
>>> IMHO we should start with openjdk 6: remove the old versions and add the
>>> latest (1.13.8)!
>>>
>>> What do you think? Any comments?
>>
>> If you are willing to work on that, I have no problem.
>
> As I mentioned, I will update openjdk-7 as long as I need it. If it's
> okay for you I will send patches for these updates to the ml. From a
> current point of view it looks like I will provide updates until end of
> 2016 at least.

Yes; fine with me.

>> The point is Java is complex and we have few people working on this so
>> I prefer one well maintained version than 3 oddly ones.
>
> I understand that. But currently I have no need for openjdk-8. Therefore
> my focus is on keeping openjdk-7 up to date.

Sure.

Specially if your app works with OpenJDK 8 I see no much benefit in
keep using the old version. Do you see it differently?

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: meta-java - recent commits break foreign (MY!) builds ...
  2015-08-31 13:06         ` Otavio Salvador
@ 2015-08-31 13:39           ` Richard Leitner
  0 siblings, 0 replies; 7+ messages in thread
From: Richard Leitner @ 2015-08-31 13:39 UTC (permalink / raw)
  To: openembedded-devel; +Cc: Otavio Salvador

On 08/31/2015 03:06 PM, Otavio Salvador wrote:
> On Mon, Aug 31, 2015 at 4:33 AM, Richard Leitner
> <richard.leitner@skidata.com> wrote:
>>> The point is Java is complex and we have few people working on this so
>>> I prefer one well maintained version than 3 oddly ones.
>>
>> I understand that. But currently I have no need for openjdk-8. Therefore
>> my focus is on keeping openjdk-7 up to date.
> 
> Sure.
> 
> Specially if your app works with OpenJDK 8 I see no much benefit in
> keep using the old version. Do you see it differently?
> 

If we have verified that everything is working as expected with
openjdk-8 we will for sure think of updating to it. But as we have a
quite big application stack and long QA/testing phases this may take
some time.
Therefore we will stay (as far as I can currently tell) at least for
another year with openjdk-7.


^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2015-08-31 13:39 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-08-24 16:48 meta-java - recent commits break foreign (MY!) builds Jens Rehsack
2015-08-27 13:28 ` Otavio Salvador
2015-08-28  6:59   ` Richard Leitner
2015-08-28 14:11     ` Otavio Salvador
2015-08-31  7:33       ` Richard Leitner
2015-08-31 13:06         ` Otavio Salvador
2015-08-31 13:39           ` Richard Leitner

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.