| [cbi-dev] CBI status as of 2012-08-08 |
|
Hi Everyone, I did some comparisons of the current CBI build as compared to the upstream platform build Juno release to understand how close we are today to the released build. It looks like CBI is still missing 15 IUs and is pulling in an extra 18 IUs which shouldn't exist. I created Bug 386884 [1] to track why we're missing IUs and Bug 386885 [2] to track why we have extra IUs. Number of IUs: CBI: 513 UPS: 510 Missing from CBI: javax.el.source org.eclipse.cvs.source org.eclipse.equinox.launcher.source org.eclipse.jdt.source org.eclipse.pde.source org.eclipse.rcp.source org.eclipse.rcp_root org.eclipse.jdt.doc.isv org.eclipse.platform.doc.isv org.eclipse.rcp.configuration_root.gtk.linux.x86_64 org.eclipse.sdk.ide toolingorg.eclipse.configuration toolingorg.eclipse.sdk.ide.config.gtk.linux.x86_64 toolingorg.eclipse.sdk.ide.configuration toolingorg.eclipse.sdk.ide.ini.gtk.linux.x86_64 I think the missing .source files are related to generated features in build.properties from org.eclipse.sdk [3]. Regarding the *.doc.isv items I think that's being worked on in Bug 385410 [4]. The org.eclipse.sdk.ide bundles I think is due to the "uid" field in the product file. We're currently using uid="org.eclipse.sdk.epp" but I guess it should be changed to org.eclipse.sdk.ide. Below is a list of bundles CBI is pulling in that does not exist in the upstream platform build. CBI pulling in but not in Upstream: javax.annotation.source javax.inject.source org.apache.batik.css.source org.apache.batik.util.gui.source org.apache.batik.util.source org.eclipse.core.net.linux.x86_64.source org.eclipse.rcp.configuration.feature.jar org.eclipse.sdk.epp org.eclipse.sdk.epp.executable.gtk.linux.x86_64 org.eclipse.sdk.epp.executable.gtk.linux.x86_64.eclipse org.w3c.css.sac.source org.w3c.dom.smil.source org.w3c.dom.svg.source toolingorg.eclipse.sdk.epp.application toolingorg.eclipse.sdk.epp.config.gtk.linux.x86_64 toolingorg.eclipse.sdk.epp.configuration toolingorg.eclipse.sdk.epp.executable.gtk.linux.x86_64 toolingorg.eclipse.sdk.epp.ini.gtk.linux.x86_64 Bundles with classes different (with patch from Bug 386646 [5]): eclipse.jdt.core:org.eclipse.jdt.compiler.apt:1.0.500-SNAPSHOT eclipse.platform.ui:org.eclipse.ui.examples.multipageeditor:3.2.200-SNAPSHOT eclipse.platform.ui:org.eclipse.ui.examples.readmetool:3.3.100-SNAPSHOT eclipse.platform.ui:org.eclipse.ui.tests.views.properties.tabbed:3.5.100-SNAPSHOT eclipse.platform.team:org.eclipse.compare.examples:3.1.200-SNAPSHOT eclipse.platform.team:org.eclipse.compare.examples.xml:3.3.200-SNAPSHOT eclipse.platform.team:org.eclipse.team.examples.filesystem:3.4.100-SNAPSHOT eclipse.platform.team:org.eclipse.team.tests.core:3.7.0-SNAPSHOT eclipse.platform.ua:org.eclipse.ua.tests:3.3.500-SNAPSHOT I've also attached status.zip which contains the diffs of the 2 builds with and without version numbers. I noticed CBI was pulling in one package org.apache.commons.logging with a newer release than what upstream Juno was pulling in. < org.apache.commons.logging=1.1.1.v201101211721 < org.apache.commons.logging.source=1.1.1.v201101211721 --- > org.apache.commons.logging=1.0.4.v201101211617 > org.apache.commons.logging.source=1.0.4.v201101211617 I was able to resolve this issue by forcing in the poms to tell the bundles to pull in org.apache.commons.logging <1.1. I'll be pushing a patch for this into CBI repos shortly. There is still a number of bundles which CBI is producing as a different version than what Upstream produced but I'm not sure if this is intentional or if it's a problem. Thanh [1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=386884 [2] https://bugs.eclipse.org/bugs/show_bug.cgi?id=386885 [3] http://dev.eclipse.org/mhonarc/lists/cbi-dev/msg00595.html [4] https://bugs.eclipse.org/bugs/show_bug.cgi?id=385410 [5] https://bugs.eclipse.org/bugs/show_bug.cgi?id=386646 |
Attachment:
status.zip
Description: Zip archive