2009/6/8 Sahoo <[email protected]> > Richard S. Hall wrote: > >> First, I am not sure whether that method only returns the bundle content >> entries...you might need to verify that and create an issue if it is >> returning fragment entries too. >> > I will try that when I have time. > >> If I remember correctly, getEntryPaths() works like typing "ls", in that >> it shows you the contents of the requested directory. Then you need to >> recurse downward for each subdirectory, which are the names ending with "/". >> > That's what the code I sent earlier tries to do. > >> If the JAR does not have directory entries, then you will not see them in >> the output, which will make it difficult to traverse. This is mentioned in >> the spec. >> > This was where I felt the spec was not very clear when I read it. The spec > does not define what "sub-path" means, does it? In the example jar file, > there is no directory by name WEB-INF/. Then, how can WEB-INF be the > sub-path of WEB-INF/web.xml? > > Are you saying the jar itself is not a valid? If it is a valid jar, how can > one iterate over all the entries? An alternative trick is OK with me as > well. >
as I understand it jars without interim directory entries are strictly not valid according to the official zip spec - however, most jar handling routines are typically lenient and will fill in missing directories where necessary in this case Felix could detect that there are entries named "WEB-INF/...etc.." and add WEB-INF to the entries at the root - I suspect this is what Equinox and Knopflerfish do Thanks, > Sahoo > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > -- Cheers, Stuart
