Thanks for the info, I now parse the filename field manually (actually using the java-mail MimeUtility.parseText method) and it now works.
You are also right about the second part: In 0.7 I get a MaximalBodyDescriptor (actually the callback still returns an almost empty BodyDescriptor which is just an interface declaration, but it's actually a MaximalBodyDescriptor ) and it has a similar method to get the filename. kind regards Michael On 09.02.2012, at 14:41, Oleg Kalnichevski wrote: > On Wed, 2012-02-08 at 11:21 +0100, Michael Kussmaul wrote: >> Hi >> >> I still use mime4j 0.6 because it had the attachment-name in the >> BodyDescriptor (as part of the callback: public void body(BodyDescriptor >> desc, InputStream stream) throws MimeException, IOException) which seems to >> be removed in 0.7? > > I am pretty sure it is still there. > >> But using the old 0.6 version I encountered, that if the filename has >> Umlaute or other special characters in it, it does not fully decode the >> file-name, e.g. an original filename: >> >> "Optimized Dällikon-Bern.gpx" >> is being reported as: >> =?iso-8859-1?Q?Optimized_D=E4llikon-Bern=2Egpx?= >> >> I was wondering what the "best practices" are to access and decode the >> filename e.g. in version 0.7.2 of mime4j? Looking at your examples >> "org.apache.james.mime4j.samples.tree.MessageTree" I only found how to use >> the DOM API, but is this also possible using your event API? >> > > One can certainly parse header fields inside #field(Field rawField) > method either using field parsers from DOM module or using a custom > parser of some sort. Personally that is how I use mime4j. > > Hope this helps > > Oleg > >> In any case many thanks for mime4j, it's a great library! >> Michael > >
