Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [tycho-user] Qualifier replaced by Timestamp that lies in the future

Since it seems to work correctly locally, maybe you should instead report a bug against the CBI product: http://wiki.eclipse.org/CBI.  Or contact that group.  I can't imagine why Tycho would behave differently locally vs. on the Eclipse Foundation's CBI.  

---Tom

On 5/4/16, 5:26 AM, "tycho-user-bounces@xxxxxxxxxxx on behalf of Patrick Bänziger" <tycho-user-bounces@xxxxxxxxxxx on behalf of Patrick.Baenziger@xxxxxxxxxxxxxxxx> wrote:

Hi Michael, thanks for your input.

 

I grepped through the involved pom files and we don’t reference the jgit timestamp provider at all.

We do explicitly configure it to use the default provider in a parent pom, and it also echoes this in the attached maven debug:

 

[DEBUG]   (f) timestampProvider = default

 

Cheers, Patrick

 

From: tycho-user-bounces@xxxxxxxxxxx [mailto:tycho-user-bounces@xxxxxxxxxxx] On Behalf Of Keppler Michael (ETAS/ESW4)
Sent: Mittwoch, 4. Mai 2016 10:51
To: Tycho user list
Subject: Re: [tycho-user] Qualifier replaced by Timestamp that lies in the future

 

Just a guess: Is the timestamp provider set to the jgit timestamp provider? If so, this might be related to one of your _commits_ (on the server) actually having a timestamp in the future?

 

Ciao, Michael

 

Von:tycho-user-bounces@xxxxxxxxxxx [mailto:tycho-user-bounces@xxxxxxxxxxx] Im Auftrag von Patrick Bänziger
Gesendet: Mittwoch, 4. Mai 2016 09:58
An: tycho-user@xxxxxxxxxxx
Betreff: [tycho-user] Qualifier replaced by Timestamp that lies in the future

 

Hi all,

 

we’re using Tycho 0.22 to build our Eclipse SDK for Eclipse Scout and encounter odd behaviour in our nightly builds.

After lots of debugging, we would be grateful if anyone has any insights…

 

 

We use the tycho-packaging-plugin to replace the qualifier part with the build timestamp, for example: 6.0.0.qualifier -> 6.0.0.20160504-0906

Locally, this builds just fine and the build-qualifier outputs a timestamp that makes sense.

Building this however on the Eclipse Foundation infrastructure, the timestamp used to replace the qualifier lies in the future!

 

This is the output from a build that finished at 2016-05-02 09:54:54 -04:00 (corrected for interleaved output by parallel build):

 

[INFO] --- tycho-packaging-plugin:0.22.0:build-qualifier-aggregator (default-build-qualifier-aggregator) @ org.eclipse.scout.sdk-feature ---

[DEBUG] Configuring mojo org.eclipse.tycho:tycho-packaging-plugin:0.22.0:build-qualifier-aggregator from plugin realm ClassRealm[plugin>org.eclipse.tycho:tycho-packaging-plugin:0.22.0, parent: sun.misc.Launcher$AppClassLoader@4e25154f]

[DEBUG] Configuring mojo 'org.eclipse.tycho:tycho-packaging-plugin:0.22.0:build-qualifier-aggregator' with basic configurator -->

[DEBUG]   (f) baseDir = /jobs/genie.scout/scout-integration-6.0-SDK-nightly/workspace/org.eclipse.scout.sdk-feature

[DEBUG]   (f) execution = org.eclipse.tycho:tycho-packaging-plugin:0.22.0:build-qualifier-aggregator {execution: default-build-qualifier-aggregator}

[DEBUG]   (s) format = yyyyMMdd-HHmm

[DEBUG]   (f) packaging = eclipse-feature

[DEBUG]   (f) project = MavenProject: org.eclipse.scout.sdk:org.eclipse.scout.sdk-feature:6.0.0-SNAPSHOT @ /jobs/genie.scout/scout-integration-6.0-SDK-nightly/workspace/org.eclipse.scout.sdk-feature/pom.xml

[DEBUG]   (f) session = org.apache.maven.execution.MavenSession@4b65bd55

[DEBUG]   (f) timestampProvider = default

[DEBUG] -- end configuration --

[INFO] The project's OSGi version is 6.0.0.20160506-0435

 

 

What we checked:

·         The time of the machine is set properly (UTC-4, but otherwise not significant drift).
We can see this as other “plain maven” build artifacts have the expected timestamp.

·         Performing the build without parallel maven build (no “-T” option) does not change it

 

We do not have remote debug access the machine. If anyone is interested, we can provide the effective POM for this part of the build.

 

Thanks in advance for any insights,

Patrick

 




Back to the top