Hi John, Thanks for the quick and clear response! I'll fix the import-package header.
I also logged the Jira: https://issues.apache.org/jira/browse/ARIES-1381 Best regards, Wouter Bancken 2015-08-14 17:17 GMT+02:00 John W Ross <[email protected]>: > To be clear, the Import-Package header is incorrect. The version is an > attribute, not a directive. It should be > >> Import-Package: >> org.objectweb.asm;version="[5.0.4,5.0. >> 4]",org.objectweb.asm.signature;version="[5.0.4,5.0. >> 4]",org.objectweb.asm.tree;version="[5.0.4,5.0.4]" > > Having said that, custom attributes and directives are allowed. In your > case, what I think should have happened is > > (1) The subsystem install should have succeeded. > (2) The version should have been treated as a custom directive. > (3) The actual package import versions would assume the default of > [0.0.0,infinity) since a version attribute was not specified. > > So yes, I think this is a bug worthy of a JIRA, although you're probably > happy to know about this right away rather than wondering why you got > wired to org.objectweb.asm 1.0. > > > >> From: Wouter Bancken <[email protected]> >> To: [email protected] >> Date: 08/14/2015 09:53 AM >> Subject: GenericDirective cannot be cast to VersionRangeAttribute >> >> Dear, >> >> I'm getting the following exception when trying to install a subsystem: >> GenericDirective cannot be cast to VersionRangeAttribute >> >> This exception is thrown when trying to parse the following >> Import-Package header: >> >> Import-Package: >> org.objectweb.asm;version:="[5.0.4,5.0. >> 4]",org.objectweb.asm.signature;version:="[5.0.4,5.0. >> 4]",org.objectweb.asm.tree;version:="[5.0.4,5.0.4]" >> >> The exception is thrown at line 155 of the ImportPackageHeader class: >> return > (VersionRangeAttribute)myParameters.get(Constants.VERSION_ATTRIBUTE); >> >> The reason for the wrong class is that the ParameterFactory class >> cannot handle the ':=' in the specification of the version. It only >> works for a single '=' without a colon: >> if (symbol.equals("=")) { >> return AttributeFactory.createAttribute(name, value); >> } >> >> Should I log a bug for this behaviour? >> >> Best regards, >> Wouter Bancken >> >
