Thanks for your quick answer.
So if I do understand from your answer and examples that I have find across several projects
I have 3 solutions, I could:
1. Add a shell command task in my Hudson job
2. Add a maven plugin (ant) in the normal build process
3. Add shell command that are deployed as a cron on the download.eclipse.org server; that is in charge to retrieve the build artifacts
Any idea what is the best approve between the 3 solutions?
It is strange that there is no official shell script that make the job for any eclipse project.
This script could define the official process in term of naming convention etc… It seems that every project has its own scripts…
Any idea if there is an official Maven/ Tycho plugin that could take all of this in order to standardize the way eclipse projects publish their artifacts?
De : cbi-dev-bounces@xxxxxxxxxxx [mailto:cbi-dev-bounces@xxxxxxxxxxx] De la part de Mickael Istria
Envoyé : lundi 26 octobre 2015 18:51
À : cbi-dev@xxxxxxxxxxx
Objet : Re: [cbi-dev] Procedure to deploy to download.eclipse.org
On 10/26/2015 06:38 PM, LE FEVRE FRANCOIS wrote:
Could you give advice / best pratices to deploy directly from the HIPP instance to the download area?
Is there any maven plugin? With a dedicated profile ?
I usually add a "Shell" step after the Maven one that takes care of copying the bits to the download area. The root of the download area in mounted in /home/data/httpd/download.eclipse.org/ .
See for example, in https://hudson.eclipse.org/swtbot/job/swtbot-tycho/
# Clean previous
rm -rf /home/data/httpd/download.eclipse.org/technology/swtbot/snapshots/*
cp -r org.eclipse.swtbot.updatesite/target/repository/* /home/data/httpd/download.eclipse.org/technology/swtbot/snapshots/
zip -r repository.zip
I know some other have implemented similar steps in a Maven profile hooked on the deploy phase and run "mvn deploy". I don't know if one is really better than the other.