On Mar 26, 2014, at 1:29 AM, Nakayama Kenjiro <nakayamakenj...@gmail.com> wrote:
> Recently I wrote new dissector by pure Lua[1] and I am thinking about > submitting a request to include the dissector to Wireshark. > But as far as I checked upstream, there are no pure Lua dissector yet. Right, none yet. The challenge I think is that most of the core developers don't know Lua nearly as well as C, so it would be hard for them to fix any bugs, or enhance it in the future to support some new version of the protocol being dissected (JGroups in your case). But as long as they're ok with the idea, I think it would be great to have such. It might even bring in more protocols, because it might be easier for some people to write one in Lua than deal with trying to figure out how to compile the source, or learn C if they don't know it already. An alternative approach is to do what many other apps do: have a mechanism to find/retrieve/install/update/remove third-party plugins from a common repository site. That way the plugins themselves aren't part of the application package and aren't officially maintained by the core application developers, but instead are maintained by the third-parties (or not, and they die off, which isn't a bad thing either). Many good text editors/IDEs do that, for example, as do browsers, etc. Some even do it with the ability to pull plugins from github directly, instead of storing the plugins in their own site. > And I don't know where shoudld I put the code in and how to modify Makefiles. Right now two Lua files are included with Wireshark: init.lua, and console.lua. Both are in epan/wslua and copied in the global configuration directory for installs/packages. (well, init.lua is actually generated from a perl script using template-init.lua) If you grep the code for "console.lua" you'll see the various files that move it around. But the global configuration directory isn't the right place really for Lua dissectors, nor is epan/wslua for the source repository. They should probably be put into the plugins/ source directory, and be installed into the global plugins directory. But we can work out those details if/when there's consensus to include Lua dissectors. For example we might wan to change the loading code a bit for plugins, to have it check if a file of the same name exists in the user's personal plugins directory, and if so then to load that one instead of the one in the global plugins dir. Or use a filename version schema and load the newest one. > So my question is > Is there any guideline to include pure Lua dissector? If no guideline yet, > why don't we make it? > > This is my personal opinion, now Lua became good for writing a new dissector > and samples like /test/lua/dissector.lua is really great. New dissector > written in Lua request will increase in the future. > > [1] https://github.com/nak3/jgroups-wireshark-dissector > > Thanks, > > Kenjiro > > -- > Kenjiro NAKAYAMA <nakayamakenj...@gmail.com> > GPG Key fingerprint = ED8F 049D E67A 727D 9A44 8E25 F44B E208 C946 5EB9 > ___________________________________________________________________________ > Sent via: Wireshark-dev mailing list <wireshark-dev@wireshark.org> > Archives: http://www.wireshark.org/lists/wireshark-dev > Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev > mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
___________________________________________________________________________ Sent via: Wireshark-dev mailing list <wireshark-dev@wireshark.org> Archives: http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
