Mark Dewey wrote: > No. It doesn't work. If it did work, then sure, I'd be serious. At the > very least it would give you to know there was a way to do it.
>From what I've read there is no portable way to determine the path to a running executable. On various OS's that support /proc, you can often glean the path from there. For example on linux, /proc/<pid#>/exe is supposed to be a symlink to the original executable. You can use readlink to determine the path. On BSD, there's a different entry that contains the actual posix path. Seems like when I've encountered people asking about this on Linux forums, invariably bigots come on and say, "why would you want to know this?" or "needing to know the path of the execution is really bad design." Whether this is true or not I cannot say. Generally the unix way is to follow certain standards for placing data files. IE if you're installed for the whole system you put data files in /usr/share or /usr/lib/yourapp/. Libraries would be in /usr/lib (or /usr/lib64) and plugins typically in /usr/lib/<yourapp>/. Configs in the home directory under a dot file (what a mess), or under the new freedesktop standards location, $HOME/.config[1]. Most programs use macro variables that are set at compile time to set these paths. Since package managers are normally used, this works very well. The idea of self-contained, relocatable packages, does not seem to hold sway in the Linux world. Maybe it should. I don't know. _______________________________________________ Vala-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/vala-list
