Community
Participate
Working Groups
Using eclipse.buildId=I20080909-1121 java.fullversion=J2RE 1.6.0 IBM J9 2.4 Windows XP x86-32 jvmwi3260-20080415_18762 (JIT enabled, AOT enabled) J9VM - 20080415_018762_lHdSMr JIT - r9_20080415_1520 GC - 20080415_AA BootLoader constants: OS=win32, ARCH=x86, WS=win32, NL=fr_CA Command-line arguments: -os win32 -ws win32 -arch x86 -consolelog -console, I got lot of errors in my workspace for many projects and the error message was: NLS missing message: BundleErrorReporter_MissingVersion in: org.eclipse.pde.internal.core.pderesources
*** Bug 246834 has been marked as a duplicate of this bug. ***
In every plug-in MANIFEST.MF, I get tons of errors with a broken message "NLS missing message: BundleErrorReporter_MissingVersion in: org.eclipse.pde.internal.core.pderesources", 1 for each package in the "Export-Package:" and "Require-Bundle:" sections. Please request a rebuild when this is fixed.
How can this can happen? Is there no self-hosting taking place on your own code? Are there no automated or at least manual smoke tests?
(In reply to comment #3) > How can this can happen? Is there no self-hosting taking place on your own > code? Are there no automated or at least manual smoke tests? > Daniel, I don't understand your comment at all! :) Perhaps you commented on a wrong bug? The issue here is "simply" that Chris forgot to commit its local 'pderesources.properties' while fixing bug 205198, I suppose
>Daniel, I don't understand your comment at all! :) Perhaps you commented on a >wrong bug? No, I did not. The point is, all users that switch to the new build are toast. Their workspace is full of errors. Committers should at least run their changes once and I would expect such failures to be detected by automated tests. Cleary, PDE people did not self-host on their newest code before RELEASING it to an I-build.
Chris, why is the default severity for a missing import "Error"? I think it should be warning or ignore at best (until the SDK decides on how it wants to deal with package versioning). We shouldn't be putting errors in the face of developers until we have tooling to fix the problems (and API tooling has no support for package version validation). This is a blocker and needs to be fixed so we have a decent I-build before the milestone.
I found the workaround to this is to set the following three severity settings to "Ignore": Preferences > Plug-in Development > Compilers > Versioning: * Missing versions on required bundles * Missing versions on exported packages * Missing versions on imported packages
For some obscure reason, the patch for bug 205198 hasn't been committed entirely (org.eclipse.pde.core has been skipped). I'll attach the patch corresponding to the diff between patch for bug 205198 and what actually is in HEAD.
morining all. I'm investigating the issue now.
Created attachment 112206 [details] patch correcting the NLS missing key issue + default level set to ignore for errors wrt versions as said in my previous comment, this patch is *just* the result of applying the patch for 205198 to HEAD @Daniel: I tested and it works :)
committed and tagged. I will issue a rebuild. I am not sure why org.eclipse.pde.core wasn't committed from my old patch in bug 205198 Thanks Ben.
Verified in N20080910-2000.
