Bugzilla – Full Text Bug Listing
| Summary: | No MANIFEST in Luna distribution | ||
|---|---|---|---|
| Product: | [Eclipse Project] Equinox | Reporter: | Ed Willink <ed> |
| Component: | p2 | Assignee: | P2 Inbox <equinox.p2-inbox> |
| Status: | CLOSED INVALID | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | ahunter.eclipse, borlander, mistria |
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Windows NT | ||
| Whiteboard: | |||
| Bug Depends on: | |||
| Bug Blocks: | 438665 | ||
|
Description
Ed Willink
bash-3.2$ jar -xvf org.eclipse.gmf.runtime.common.ui.services.action_1.7.0.201406111639.jar inflated: META-INF/MANIFEST.MF inflated: META-INF/ECLIPSE_.SF inflated: META-INF/ECLIPSE_.RSA created: META-INF/ I'm not sure what your comment means. It's certainly hard to undertstand how a manifest could be missing and I successfully installed GMF Runtime a week ago. Bug 438665 is the first indication that sometibng is awry. My observation for this bug was based on logging on to build.eclipse.org and vi'ing the relevant jar to inspect the contents. (A bit frightening that I can actually use vi on a releases JAR, but I remembered to :quit.) >> I'm not sure what your comment means.
Hum, being really surprised by the bug description, I
- downloaded the jar from the link above,
- un-jarred it,
- found the MANIFEST.MF inside,
- posted the first few lines of the log, showing the successful extraction of the MANIFEST.MF here
I see more details in the #438665, and there is indeed some problem, but the claim that "there is no MANIFEST.MF in the archive" is wrong.
I am not seeing this. I installed the GMF Runtime and Examples from http://download.eclipse.org/releases/luna and everything is working. Is this a post Luna GA repo? It should be the normal GA. I tried (GMF Runtime and Examples); still failed. I restarted Eclipse; still fails. I rebooted Windows; still fails. (NB original report was GMF Runtime). What is the exception you are getting? How do you know it is the one bundle? (In reply to Anthony Hunter from comment #6) > What is the exception you are getting? How do you know it is the one bundle? See Bug 438665 (In reply to Ed Willink from comment #7) > (In reply to Anthony Hunter from comment #6) > > What is the exception you are getting? How do you know it is the one bundle? > > See Bug 438665 All I can see there is "Attempting to install GMF Runtime direct from P2". Are you installing from a zip or are your installing from a update site? If it is an update site then which one is it? The Luna update site at http://download.eclipse.org/releases/luna works fine. I am not seeing any other complaints, and GMF is in the modelling package for Luna and there were no complaints building and users running with that. And GMF is installed into the eclipse runtime so that the Junit tests could run, and there are tests for common.ui.services.action so we know the bundle is loading. I am stumped as to where an issue could be. It is the http://download.eclipse.org/releases/luna site that fails repeatedly for me. I was second on Bug 438665. It's certainly odd, but as yet I have failed to install Papyrus on my laptop. (In reply to Ed Willink from comment #9) > It's certainly odd, but as yet I have failed to install Papyrus on my laptop. I tried an installation from scratch and it worked. So the problem is a bad configuration (possibly workspace) cache. The new installation initially reported that it could not find http://download.eclipse.org/releases/luna and so had to be restarted to find it. I've seen this before. I have also reported that about 10% of my attempts to start a debugger report a VM connection failure long before the connection timeout. I also observe that GIT fails to connect to a repo on about 10% of attempts. Iy seems that either Eclipse has acquired some flaky I/O component that ripples downstream or my laptop has a wierd problem. The issue described here doesn't seem to be reproducible in general and doesn't seem so related to p2. Feel free to reopen and re-organize subject if you think there is still something p2 should fix on this topic. |