Possible bug in feature repository version ranges

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Possible bug in feature repository version ranges

Martin Nielsen
Hi everyone
My and my colleagues found a possible bug when resolving feature repo
version ranges. The bug seems to be present on Windows machines, but does
not seem to occur on debian(And likely any linux distro).

I have checked out Karaf and replicated the bug here:

https://github.com/mnybon/karaf/tree/versionrange_windows

The bug seems to be found in the karaf-maven-plugin karaf assembly and
happens when a version range is found in a feature file set to be installed
into the system dir. I have used the new CXF release which has a pax cdi
repository import with version range. On our linux server, everything seems
to work correctly, but on windows, the feature with the version range is
placed in system with the version range intact in the name.

On windows the feature is placed under:
karaf\assemblies\apache-karaf-minimal\target\assembly\system\org\ops4j\pax\cdi\pax-cdi-features\[1.0.0.RC1,2)
On debian the feature is placed under:
karaf/assemblies/apache-karaf-minimal/target/assembly/system/org/ops4j/pax/cdi/pax-cdi-features/1.0.0.RC1

Note pax-cdi-features\*[1.0.0.RC1,2)* vs pax-cdi-features/*1.0.0.RC1*

I am using the same version of maven and Java 8 on both machines, so I
believe the error is contained to windows machines.
I have also had a few colleagues test it on their windows machines, with
different maven and java versions, the error is constant and reproduced
consistently.


The created karaf assembly does not exactly appreciate the
*[1.0.0.RC1,2)* version
when it starts up. So it basically renders builds from windows machines
useless.

-Martin
Reply | Threaded
Open this post in threaded view
|

Re: Possible bug in feature repository version ranges

jbonofre
Hi Martin,

thanks for the report. Seems like a Windows environment issue. Let me try to
reproduce on a VM.

Thanks again,
Regards
JB

On 01/06/2018 12:07 PM, Martin Nielsen wrote:

> Hi everyone
> My and my colleagues found a possible bug when resolving feature repo
> version ranges. The bug seems to be present on Windows machines, but does
> not seem to occur on debian(And likely any linux distro).
>
> I have checked out Karaf and replicated the bug here:
>
> https://github.com/mnybon/karaf/tree/versionrange_windows
>
> The bug seems to be found in the karaf-maven-plugin karaf assembly and
> happens when a version range is found in a feature file set to be installed
> into the system dir. I have used the new CXF release which has a pax cdi
> repository import with version range. On our linux server, everything seems
> to work correctly, but on windows, the feature with the version range is
> placed in system with the version range intact in the name.
>
> On windows the feature is placed under:
> karaf\assemblies\apache-karaf-minimal\target\assembly\system\org\ops4j\pax\cdi\pax-cdi-features\[1.0.0.RC1,2)
> On debian the feature is placed under:
> karaf/assemblies/apache-karaf-minimal/target/assembly/system/org/ops4j/pax/cdi/pax-cdi-features/1.0.0.RC1
>
> Note pax-cdi-features\*[1.0.0.RC1,2)* vs pax-cdi-features/*1.0.0.RC1*
>
> I am using the same version of maven and Java 8 on both machines, so I
> believe the error is contained to windows machines.
> I have also had a few colleagues test it on their windows machines, with
> different maven and java versions, the error is constant and reproduced
> consistently.
>
>
> The created karaf assembly does not exactly appreciate the
> *[1.0.0.RC1,2)* version
> when it starts up. So it basically renders builds from windows machines
> useless.
>
> -Martin
>

--
Jean-Baptiste Onofré
[hidden email]
http://blog.nanthrax.net
Talend - http://www.talend.com
Reply | Threaded
Open this post in threaded view
|

Re: Possible bug in feature repository version ranges

Martin Nielsen
Hello Jean-Baptiste.
Did you manage to recreate the bug?
If not i could attempt to dig a little deeper, but i assumed it was every
windows machine that had the problem.

-Martin

On Sat, Jan 6, 2018 at 2:48 PM, Jean-Baptiste Onofré <[hidden email]>
wrote:

> Hi Martin,
>
> thanks for the report. Seems like a Windows environment issue. Let me try
> to reproduce on a VM.
>
> Thanks again,
> Regards
> JB
>
> On 01/06/2018 12:07 PM, Martin Nielsen wrote:
>
>> Hi everyone
>> My and my colleagues found a possible bug when resolving feature repo
>> version ranges. The bug seems to be present on Windows machines, but does
>> not seem to occur on debian(And likely any linux distro).
>>
>> I have checked out Karaf and replicated the bug here:
>>
>> https://github.com/mnybon/karaf/tree/versionrange_windows
>>
>> The bug seems to be found in the karaf-maven-plugin karaf assembly and
>> happens when a version range is found in a feature file set to be
>> installed
>> into the system dir. I have used the new CXF release which has a pax cdi
>> repository import with version range. On our linux server, everything
>> seems
>> to work correctly, but on windows, the feature with the version range is
>> placed in system with the version range intact in the name.
>>
>> On windows the feature is placed under:
>> karaf\assemblies\apache-karaf-minimal\target\assembly\system
>> \org\ops4j\pax\cdi\pax-cdi-features\[1.0.0.RC1,2)
>> On debian the feature is placed under:
>> karaf/assemblies/apache-karaf-minimal/target/assembly/system
>> /org/ops4j/pax/cdi/pax-cdi-features/1.0.0.RC1
>>
>> Note pax-cdi-features\*[1.0.0.RC1,2)* vs pax-cdi-features/*1.0.0.RC1*
>>
>> I am using the same version of maven and Java 8 on both machines, so I
>> believe the error is contained to windows machines.
>> I have also had a few colleagues test it on their windows machines, with
>> different maven and java versions, the error is constant and reproduced
>> consistently.
>>
>>
>> The created karaf assembly does not exactly appreciate the
>> *[1.0.0.RC1,2)* version
>> when it starts up. So it basically renders builds from windows machines
>> useless.
>>
>> -Martin
>>
>>
> --
> Jean-Baptiste Onofré
> [hidden email]
> http://blog.nanthrax.net
> Talend - http://www.talend.com
>