I think Jörg sums it up quite nicely. But there is also the issue about being lenient about what we accept and strict about what we write. So I think we might want to be keeping our eyes on the consequences of this change; if it turns out to be enough of a problem we can consider adding some pre-parsing to the jdk manifest class usage.
But I would *prefer* to keep it as-is, since that means we don't add any bugs on top of the JDK logic ;) Kristian 2012/2/3 Jörg Schaible <[email protected]>: > Hi Jane, > > Jane Young wrote: > >> Hi, >> >> There is a new maven-jar-plugin 2.4 and maven-archiver 2.5 released on >> Feb 1st. >> >> I have a MANIFEST file that contains a very long classpath. When >> executing Maven 3 with maven-jar-plugin 2.4 and maven-archiver 2.5, it >> will fail with the following message: >> >> [ERROR] Failed to execute goal >> org.apache.maven.plugins:maven-jar-plugin:2.4:jar (default-jar) on >> project simple-test: Error assembling JAR: Unable to read manifest file >> (line too long) -> [Help 1] >> [ERROR] >> >> This is working with previous versions of maven-jar-plugin but it's >> failing with maven-jar-plugin 2.4. >> >> Attached is a simple test. Can someone verify if this is bug? > > This is *no* bug, *your* MANIFEST is invalid (see manifest specification in > JDK). The difference is that Maven uses now the JDK to read the manifest, > while it did it in earlier versions on its own. So, don't expect that your > jar file is working anywhere. > > - Jörg > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
