Martin Probst wrote: > Charles: where can I find you're patches to JLine, in the sourceforge > bug tracker?
Yes, there, and probably a few in our tracker somewhere as well. I think we have a contrib that's about to fork JLine though because we're not getting responses for new releases. Perhaps if the original JLine maintainer has disappeared we should fork it together and start spinning releases again? > As for the alternatives, I think what Per writes is certainly doable, > but it sounds a bit complicated to me... also, I would be heavily > interested in an as much as possible pure Java solution, if only for > ease of development. Using a JNI interface to readline would probably > work and make deployment ok, but I'm not sure of the license > consequences - GPL is probably not acceptable for many users. > > Or did you mean using a JNI interface to directly call the respective > syscalls to read single keystrokes? The JLine approach of using a dll > on Windows and stty calls on UNIXes seems to work, as far as I can > see. I mean using something like this (which ships with JRuby) to wire in JLine: http://blog.headius.com/2008/10/ffi-for-ruby-now-available.html (seems to be down right now...which means Blogger's busted) Think of it as JNA for Ruby, but the general idea is that we're expanding use of an FFI/DL layer to support most of the native libs Rubyists depend on. So we only ship a single dynamic library for all supported platforms and go through that. > JLine will probably never be absolutely readline compatible, but I'd > personally prefer a Java solution which I can easily hook into my Java > project and fiddle with through extension etc. instead of a C native > library or telnet style setup. Well I guess my question is what does readline actually *do*? If we can implement the same without native calls, or make the native calls ourselves using JNA or similar, it would be even better than depending on libreadline and worrying about licensing. > Has anyone tried how/if a combination of JLine (or something similar) > and Nailgun works? It would need readline in the client; readline in the server wouldn't work. I think Nailgun needs to be taken over as well, though...I've wanted to add signal support, for example (^C and friends passing through to server), but it's unlikely to get in without help. > My personal wish for this would be that someone (Sun?) turns up and > provides a repository system like a cross between Maven, Apt, and OSGi > together with the JVM, a proper isolation of apps within one JVM > including a Thread.kill, and building on that an improved nailgun. The bits to do this are already out there; it may be up to us to integrate them well. Perhaps we should start a project that pulls in or forks each of these projects and integrates them more tightly? - Charlie --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "JVM Languages" group. To post to this group, send email to jvm-languages@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/jvm-languages?hl=en -~----------~----~----~----~------~----~------~--~---
