Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [orbit-dev] Backward version of org.apache.commons.codec 1.3.0 in S20100508202910?


I've opened bug 313403
https://bugs.eclipse.org/bugs/show_bug.cgi?id=313403
to track this "baseLocation" issue, but ...

The fact that we were picking up something from the baseLocation (the base builder, in this case) is a sign of a problem in versioning. I think it would be in all cases, but is for sure in this case.

It seems if there is a qualification scheme "in use" in bundles that exist already, we have to continue to follow that scheme, and make sure we always increment from there.
I don't know the whole history of this bundle, but that a version 1.3.0.v20080530-1600  exists "in the wild" means that we can't change to use a scheme that uses qualifiers
like v1_xxx ... it is immediately less than anything versioned with a year. Well, a year this century. :)

So, unsetting the baselocation might be useful, just to make sure similar errors are not occurring, but if they are, that implies we'd have other qualifiers to fix.

Thanks so much Andrew and DJ for digging in and getting to the bottom of this.





From: Andrew Niefer <aniefer@xxxxxxxxxx>
To: Orbit Developer discussion <orbit-dev@xxxxxxxxxxx>
Date: 05/18/2010 11:14 AM
Subject: Re: [orbit-dev] Backward version of org.apache.commons.codec 1.3.0 in        S20100508202910?
Sent by: orbit-dev-bounces@xxxxxxxxxxx





This is because the build is using a "baseLocation" which contains apache.commons.codec v20080530-1600, which is a higher version that the v1_3_v20100508-2020 being built. So the osgi resolver is prefering the higher version bundle.

I would suggest that the orbit build likely does not need a baseLocation and can set that to be an empty string ("baseLocation=").

-Andrew

Inactive hide details for DJ Houghton---05/18/2010 06:58:01 AM---I have no idea. The map file (and directory.txt) on the file sDJ Houghton---05/18/2010 06:58:01 AM---I have no idea. The map file (and directory.txt) on the file system looks ok but the build is produc

From:

DJ Houghton/Ottawa/IBM@IBMCA

To:

Orbit Developer discussion <orbit-dev@xxxxxxxxxxx>

Date:

05/18/2010 06:58 AM

Subject:

Re: [orbit-dev] Backward version of org.apache.commons.codec 1.3.0 in S20100508202910?

Sent by:

orbit-dev-bounces@xxxxxxxxxxx





I have no idea. The map file (and directory.txt) on the file system looks ok but the build is producing old artifacts. I'm not sure where it is getting them from. I have retagged and kicked off another build.

dj


Inactive hide details for David M Williams ---05/18/2010 12:18:55 AM---No, not intentional.  FWIW, this was touched as part of David M Williams ---05/18/2010 12:18:55 AM---No, not intentional. FWIW, this was touched as part of bug 312168

From:

David M Williams <david_williams@xxxxxxxxxx>

To:

Orbit Developer discussion <orbit-dev@xxxxxxxxxxx>

Date:

05/18/2010 12:18 AM

Subject:

Re: [orbit-dev] Backward version of org.apache.commons.codec 1.3.0 in S20100508202910?

Sent by:

orbit-dev-bounces@xxxxxxxxxxx






No, not intentional.

FWIW, this was touched as part of bug 312168

https://bugs.eclipse.org/bugs/show_bug.cgi?id=312168

But, the map file says the qualifier should be
v1_3_v20100508-2020
and the previous one should have been
v1_3_v20100427-1100

according to map file.

Therefore, I'm totally confused about how this one works ... maybe DJ can shed some light?

No matter what the cause, seems a fix is in order, so the qualifier is increased.



From: "Xiaoying Gu" <xgu@xxxxxxxxxxx>
To: "Orbit Developer discussion" <orbit-dev@xxxxxxxxxxx>
Date: 05/17/2010 11:50 PM
Subject: [orbit-dev] Backward version of org.apache.commons.codec 1.3.0 in S20100508202910?
Sent by: orbit-dev-bounces@xxxxxxxxxxx






Hi ,


I found that the org.apache.commons.codec 1.3.0 bundle in Orbit S20100508202910 (which is for Helios RC1) is versioned as 1.3.0.v20080530-1600.

But in the previous S-build its version was 1.3.0.v20100106-1700.


Is the backward versioning the intentional change?



Thanks,

Xiaoying

_______________________________________________
orbit-dev mailing list
orbit-dev@xxxxxxxxxxx

https://dev.eclipse.org/mailman/listinfo/orbit-dev

_______________________________________________
orbit-dev mailing list
orbit-dev@xxxxxxxxxxx

https://dev.eclipse.org/mailman/listinfo/orbit-dev

_______________________________________________
orbit-dev mailing list
orbit-dev@xxxxxxxxxxx

https://dev.eclipse.org/mailman/listinfo/orbit-dev

_______________________________________________
orbit-dev mailing list
orbit-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/orbit-dev



Back to the top