Bugzilla – Full Text Bug Listing
| Summary: | [CBI] Platform binary distribution (still) contains pack.gz files | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Eclipse Project] Platform | Reporter: | David Williams <david_williams> | ||||
| Component: | Releng | Assignee: | David Williams <david_williams> | ||||
| Status: | VERIFIED FIXED | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | P3 | CC: | pwebster | ||||
| Version: | 4.2.1 | ||||||
| Target Milestone: | 4.3 RC1 | ||||||
| Hardware: | PC | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
|
Description
David Williams
Just to document details, (fixing this will be a bit trial and error), there are three places in our aggregator that specify <includePackedArtifacts> Two in profiles in eclipse parent: eclipse-pack and eclipse-sign profile. One of those was false and one of those was true. I think they both need to to be false. The profile eclipse-pack is still important, since that is also what does the "pre-conditioning", apparently. Then there's the one in in eclipse.platform.repository that says "true" as is intended, so the final, main repo will be ready for "web deployment". I'll try this fix, (changing the other in eclipse-parent to false) http://git.eclipse.org/c/platform/eclipse.platform.releng.aggregator.git/commit/?id=6ad8c84ff3002909173ef25719b96decca053b53 And see if that fixes the "binary platform distribution" case ... without effecting anything else. This will take an I build to test. I can likely run a build machine only test build to get a sense if its working. Not fixed in M7 ... will need to investigate. I noticed issue comparing "the last PDE build" (where it is only 100 MB) with the one from M7 (where it is nearly 150M) http://git.eclipse.org/c/platform/eclipse.platform.releng.aggregator.git/commit/?id=cee3342e120862eb769a845c0fd423149c3851f0 I'll attempt to fix by adding <includePackedArtifacts>false</includePackedArtifacts> to all the target-platform-configuration for "products" in tychoeclipse builder. Except for those where it is not a problem. Not sure why some seem to inherit right value from eclipse parent, and other do not ... but ... even if that's it some subtle bug in Tycho, at this point best to fix just by setting explicitly where needed. That didn't fix it, either. But, see 407130 for some info that might explain why (In reply to comment #5) > That didn't fix it, either. But, see 407130 for some info that might explain > why That is, we currently create the repos "slice" from the main, final, repo, but I can't help wonder why we don't create (copy) the one that's already created at /opt/public/eclipse/builds/4I/gitCache/eclipse.platform.releng.aggregator/eclipse.platform.releng.tychoeclipsebuilder/platform/target Created attachment 230547 [details] patch to put p2.remove in ant script that creates this repo This patch contains 2 fixes. One for this bug, one for bug 407314. Paul, any comments? http://git.eclipse.org/c/platform/eclipse.platform.releng.aggregator.git/commit/?id=721bfe0dbf0cb0f3ddd48283cd170d9cdc2800f8 I think this fix is pretty straightforward, (I forgot one p2.remove, in patch). But this commit is only for this bug, not bug 407314. Verified in build I20130507-1100 ... back down to 100 Meg. |