| SVN update during materialization [message #646901] |
Sun, 02 January 2011 20:56  |
 Zoltan Ujhelyi
Messages: 392
Registered: July 2015 |
Senior Member |
|
|
Hi,
thanks for the Buckminster tool, it is really useful for build automation.
I try to use Buckminster for continous integration together with Hudson, and selected Buckminster to download the required projects for the repository (in truth, we have a somewhat messy set of repositories, but the rmap allowed us to handle this with a simple set of rules - on the other hand manually specifying where to download what would be hard, and simply updating everything downloads a lot of code, and the rmap would be still hard to write...).
However, if the code in the repository changes, Buckminster does not update its workspace, but uses the previously checked-out projects. I found this old thread that suggest providing Advisor nodes in the cquery to prevent resolving from the workspace, and that would issue an update: http://dev.eclipse.org/mhonarc/lists/buckminster-dev/msg0071 2.html Unluckily, this approach does not seem working for me.
My cquery looks like this:
<?xml version="1.0" encoding="UTF-8"?>
<cq:componentQuery xmlns:cq="http://www.eclipse.org/buckminster/CQuery-1.0" resourceMap=" https://svn.inf.mit.bme.hu/viatra/viatra_extra/R3/releng/tru nk/org.eclipse.viatra2.releng/viatra.rmap" properties=" https://svn.inf.mit.bme.hu/viatra/viatra_extra/R3/releng/tru nk/org.eclipse.viatra2.releng/buckminster.properties">
<cq:rootRequest name="org.eclipse.viatra2.releng" componentType="eclipse.feature"/>
<cq:property key="target.arch" value="*"/>
<cq:property key="target.os" value="*"/>
<cq:property key="target.ws" value="*"/>
<cq:advisorNode namePattern="^hu\.optxware\.emf2viatra\.core" useMaterialization="false" useWorkspace="false"/>
<cq:advisorNode namePattern="^hu\.bme\.mit([\.\-].+)?" useMaterialization="false" useWorkspace="false"/>
<cq:advisorNode namePattern="^hu\.optxware\.emf2viatra\.core" useMaterialization="false" useWorkspace="false"/>
</cq:componentQuery>
The referenced rmap uses the following locators:
<rm:locator pattern="^org\.eclipse\.viatra2([\.\-].+)?" searchPathRef="VIATRA_SVN" failOnError="false"/>
<rm:locator pattern="^hu\.bme\.mit([\.\-].+)?" searchPathRef="VIATRA_SVN"/>
<rm:locator pattern="^org\.eclipse\..*" searchPathRef="eclipse.org"/>
<rm:locator pattern="^hu\.optxware\.emf2viatra\.core" searchPathRef="VIATRA_SVN"/>
<rm:locator pattern="^com\.ibm([\.\-].+)" searchPathRef="eclipse.org"/>
<rm:locator pattern="^org\.apache([\.\-].+)" searchPathRef="eclipse.org"/>
<rm:locator pattern="^org\.sat4j([\.\-].+)" searchPathRef="eclipse.org"/>
<rm:locator pattern="^org\.junit([\.\-].+)" searchPathRef="eclipse.org"/>
<rm:locator pattern="^org\.junit" searchPathRef="eclipse.org"/>
<rm:locator pattern="^org\.hamcrest([\.\-].+)" searchPathRef="eclipse.org"/>
where eclipse.org points to the Helios update site, while VIATRA_SVN marks a set of repository URLs to check for the various projects. Name patterns are the same in both the locators and the cquery advisor nodes, and to projects get downloaded once.
Could you please help me to give a hint what have I done wrong, how to provide this functionality? I'm using the latest stable build of Buckminster (3.6.1) both in the IDE and in headless mode in the server.
Thank you in advance,
Zoltán Ujhelyi
|
|
|
|
|
Powered by
FUDForum. Page generated in 0.03538 seconds
] 
Search
Help
Register
Login
Home
