| Summary: | Ganymede M5 contains obsolete files and invalid *.pack.gz files | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Community | Reporter: | Martin Oberhuber <mober.at+eclipse> | ||||
| Component: | Cross-Project | Assignee: | Bjorn Freeman-Benson <bjorn.freeman-benson> | ||||
| Status: | RESOLVED FIXED | QA Contact: | |||||
| Severity: | major | ||||||
| Priority: | P2 | CC: | aniefer, bjorn.freeman-benson, denis.roy | ||||
| Version: | unspecified | ||||||
| Target Milestone: | --- | ||||||
| Hardware: | PC | ||||||
| OS: | Windows XP | ||||||
| URL: | http://download.eclipse.org/releases/ganymede | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
|
Description
Martin Oberhuber
FWIW, I ran the siteOptimizer on dsdp.eclipse.org to generate the bad pack.gz files, and uploaded the result to http://download.eclipse.org/dsdp/tm/div/ganymede_m5_fixpack.zip The archive is 43MB, and can be extracted onto releases/ganymede, it will overwrite only the 0-byte-length .pack.gz files as well as fix the invalid digest (see bug 221993). So this would be a quickfix for now, and I verified on Windows that the resulting ganymede can be installed completely. If anybody wants to have a complete archived ganymede site for local installs, it's now available on http://download.eclipse.org/dsdp/tm/div/ganymede_m5.zip Size is 298MB -- 12MB larger than the archive I had created on 25-Feb, because this time I chose not to pack any jar files that were not packed on releases/ganymede, just to be safe. This means that ECF, Mylyn, STP as wellas most of EMFT are not packed in this version. More observations, compared to the 25-Feb version: * Buckminster core version bump from 1.0.0.r08951 to 1.0.0.r08952 * Buckminster cvs version bump from 1.0.0.r08949 to 1.0.0.r08953 * net.sourceforge.lpg.lpgjavaruntime.source was Jarred, now extracted (1) and (3) - I deleted the whole staging directory and started over. (2) is due to bug 187396 (In reply to comment #2) > (2) is due to bug 187396 I'm confused... I'm running the same command as in bug 187396 comment 18 and it works for me on Ganymede, since I'm using Eclipse-3.4M5 as the builder. I thought that the Ganymatic also uses a recent Eclipse as the builder? Bug 187396 seems to be relevant for Europa only. (In reply to comment #3) I only update the Ganymatic (and its underlying Eclipse) when necessary: I prefer to keep a stable build system and not run on the bleeding edge. However, at your urging, I have updated the Eclipse that I use to run the siteOptimizer to 3.4M5. When it runs, I get errors like this: Running on /home/data/users/bfreeman/downloads/releases/ganymede/staging/features/org.eclipse.rse.ftp.source_3.0.0.v20080212-79378_kE77a9C9E535C.jar Running Pack on /home/data/users/bfreeman/downloads/releases/ganymede/staging/features/org.eclipse.rse.ftp.source_3.0.0.v20080212-79378_kE77a9C9E535C.jar STDERR: Exception in thread "main" java.lang.SecurityException: SHA1 digest error for META-INF/eclipse.inf STDERR: at sun.security.util.ManifestEntryVerifier.verify(ManifestEntryVerifier.java:253) STDERR: at java.util.jar.JarVerifier.processEntry(JarVerifier.java:225) STDERR: at java.util.jar.JarVerifier.update(JarVerifier.java:212) STDERR: at java.util.jar.JarVerifier$VerifierStream.read(JarVerifier.java:438) STDERR: at java.io.InputStream.read(InputStream.java:112) STDERR: at com.sun.java.util.jar.pack.Package$File.readFrom(Package.java:789) STDERR: at com.sun.java.util.jar.pack.PackerImpl$DoPack.readFile(PackerImpl.java:529) STDERR: at com.sun.java.util.jar.pack.PackerImpl$DoPack.run(PackerImpl.java:490) STDERR: at com.sun.java.util.jar.pack.PackerImpl.pack(PackerImpl.java:91) STDERR: at com.sun.java.util.jar.pack.Driver.main(Driver.java:279) Error: 1 was returned from command: /opt/public/ganymede/apps/ibm-java2-ppc-50/jre/bin/pack200 -E4 /home/data/httpd/download.eclipse.org/releases/ganymede/staging/features/org.eclipse.rse.ftp.source_3.0.0.v20080212-79378_kE77a9C9E535C.jar.pack.gz /home/data/httpd/download.eclipse.org/releases/ganymede/staging/features/temp.org.eclipse.rse.ftp.source_3.0.0.v20080212-79378_kE77a9C9E5 Additionally (or perhaps it is the same issue), I get zero length files, e.g., -rw-rw-r-- 1 bfreeman callistoadmin 0 2008-03-14 18:46 plugins/org.eclipse.gmf.runtime.gef.ui_1.1.0.v20080114-1111.jar.pack.gz The command line is: /shared/ganymede/apps/ibm-java2-ppc-50/bin/java -jar /home/data/users/bfreeman/ganymede/eclipse-34M5/plugins/org.eclipse.equinox.launcher_1.0.100.v20080205.jar -application org.eclipse.update.core.siteOptimizer -jarProcessor -pack -verbose -outputDir /home/data/users/bfreeman/downloads/releases/ganymede/staging/ /home/data/users/bfreeman/downloads/releases/ganymede/staging/ Hm.. it works ok for me on x86 / Sun Jvm. Are you running the latest IBM JVM? See bug 220943. Does the error also occur when running the "pack200" command verbatim as specified on the commandline? If you want to try on an x86 vserver yourself, my latest mirroring / packing / digesting script is in http://download.eclipse.org/dsdp/tm/div/mirrorGanymede.tar (In reply to comment #5) > Are you running the latest IBM JVM? Yes. See the complete log of a run: http://build.eclipse.org/ganymede/temp.log Given that this "sort of works" now (latest JVM, latest Eclipse 3.4) (modulo the fact that the IBM PPC JVM doesn't seem to work with pack200) and the fact that I'm not planning to pack200 every staging build (the pack200 step takes over an hour and thus each Ganymede build would take over an hour longer if we did that), I'm closing this bug. |