| Summary: | ptp not in latest staging repo | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Community | Reporter: | David Williams <david_williams> | ||||
| Component: | Cross-Project | Assignee: | Cross-Project issues <cross-project.inbox> | ||||
| Status: | RESOLVED FIXED | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | P3 | CC: | beth, g.watson, vivkong | ||||
| Version: | unspecified | ||||||
| Target Milestone: | --- | ||||||
| Hardware: | PC | ||||||
| OS: | Windows 7 | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
|
Description
David Williams
Looks like we're in it now. Go figure... I see it too, in staging, using p2. Why don't I see it in log? Must be searching for wrong thing? Or browser search isn't working? I searched for things like org.eclipse.ptp but will close since I guess original problem isn't a problem. It's not in the features/plugins directory though, so the install barfs. I have a hunch ... have you recently added a requirement on Java 1.6, specifically? Not that I'm aware of. All ptp components should be 1.5 only. I wonder if it was just a timing issue where the aggregator started after I uploaded to our repository? Unfortunately it looks like the last build failed, so we can't tell yet. Even if build fails as a whole due to "can not mirror" one artifact, but rest of the build continues, and I still don't see it in log of failed build. I look at the repo with normal p2 install, and do see one feature that has "featureName" as the name ... as if it is getting only the NL key, or something. Normally that should not be a fatal problem, but am wondering if that uncovers a bug in the builder, causing the silent failure? Similarly, it has key-only copyright, license, and description. It says the feature is org.eclipse.ptp.master.feature.group (or, you'd probably know it as org.eclipse.ptp.master). Is it supposed to be in your repo? perhaps you could correct that, since we don't know what else to do? Created attachment 171578 [details]
screen shot showing the 'featureName' feature
I think (hope) I fixed this in the latest build. But in any case, this would have been there for some time and other builds worked... I just checked and the master feature is still there. I'll remove it now. I do see this "1.6 jre" line in your content.xml file. Not sure how it can get there ... maybe a "product definition"? <unit id='a.jre' version='1.6.0' singleton='false'> And, not sure if it is related or not to this problem. But ... if somehow it is saying "1.6. jre is required" for things in this repo then that might be why builder fails silently ... the builder is just looking for "1.5 compatible features" (it will of course pick up 1.6 required bundles/fragments. I'm picking up on this, because I do not see this in other content.xml files I look in. Do you have some p2.inf file that could be contributing some "touchpoint"? And ... are you contributing a _product_? Or just collection of features? I seem to recall "products" take something special. I think it worked this time. I'm a bit hesitant to change anything at this late stage, but it seems like something we should definitely look into. Yes, even though overall build failed, at least I see it in log now ...
How did you fix it? :)
[exec] Mirroring meta-data from from http://download.eclipse.org/tools/ptp/updates/helios
[exec] - mirroring artifacts reference http://download.eclipse.org/tools/ptp/updates/helios
[exec] - mirroring meta-data reference http://download.eclipse.org/tools/ptp/updates/helios
[exec] Mirroring artifacts from from http://download.eclipse.org/tools/ptp/updates/helios
[exec] - mirroring artifact osgi.bundle,org.eclipse.ptp.rm.pbs.core,4.0.0.201006091432
[exec] doing copy of canonical artifact
[exec] - mirroring artifact osgi.bundle,org.eclipse.ptp.etfw.feedback,2.0.0.201006091432
[exec] doing copy of canonical artifact
[exec] - mirroring artifact osgi.bundle,org.eclipse.ptp.pldt.help,4.0.0.201006091432
[exec] doing copy of canonical artifact
[exec] - mirroring artifact osgi.bundle,org.eclipse.ptp.rdt.xlc,1.0.0.201006091432
[...]
Left it alone and prayed :-) I've opened bug 316401 to document the "silent failure" ... even though not sure we can reproduce it. |