Actually thunarx-python is nearly abandoned yet still being used by some important projects like rabbitvcs, mega, dropbox, insync alongside a few more and dont forget that maybe there are undiscovered security problem. The problem recide in that either this plugin is gtk2 and python2 and with thunar migration to gtk3 this plugin could not work properly AND python2 in 3 years will be EOL so to prevent this problem a solution could either update the plugin or merge it into thunar. mergin into thunar have the adventage that it can be more closely integrated into thunar and will have more maintainance than now. Alternatively you can update the binding to move to gtk3 and python3. Or maybe extend thunarx-2 to provide all the functionallyty that the binding do if possible.
It looks like dropbox and mega use the C extension framework and are not dependent on thunarx-python. Still, you are right that a change needs to be made. I'll work on updating thunarx-python to Gtk3/python3. I wouldn't have a problem integrating the python extension framework into Thunar. In some ways it makes sense, but I'll leave that to the Thunar maintainers to decide.
Thanks for the effort. I was looking and at least for rabbitcvs the python2 is a blocker since the filebrowser-plugins need python3 and they dont have the manpower for keep the 2 toolkits. I think we should add thunar maintainer to the cc since the binding is for thunar and to know they opinion about merging it or not.
Disclaimer: I'm not Thunar's maintainer, just a dude helping on porting it. I'm not against merging thunarx-python into thunar, but looking at others components (libs, panel...) they don't include bindings in their source tree, python bindings are in pyxfce and vala bindings in xfce4-vala, probably also outdated. It makes more sense IMHO to merge thunarx-python into pyxfce. Either way, it shouldn't make much difference, why not keep the project standalone as it currently?
pyxfce has been unmaintained for even longer than thunarx-pyhton yet I don't know any project using it and at this point I think (maybe wrong so anyone correct me) that all the functionallity used to be probided by pyxfce has landed on the rest of xfce as part of the normal program without need of bindings, same for xfce4-vala but without the landing. I think at least is easier to mainatin thunarx-python as either separated or updated since look like was the only binding that ended being used in any system. I think at this point we need the opinion of the pyxfce, thunar and thunar-pyhton about the issue.
I think it makes fine sense to keep thunarx-python the way it is. It's a pretty small codebase and as long as it works with Thunar there isn't much that needs to be done. It is feature complete, so the only changes that need to be made are bug fixes and keeping up with breakinng changes in Thunar.
Alright then, let's keep it as is. By the way, there will be breakings until Thunar 1.7.0 gets released, for example, GtkAction is deprecated and we will need to figure out a replacement.
As of v0.4.0 I've updated thunarx-python to work with Gtk+3 and the newest Thunar v1.7.x