On Jan 23, 2008 2:26 AM, Philip Ganchev <[EMAIL PROTECTED]> wrote: > But for the rest of the proposal, is there enough reason to do it? > First, it may be more humane to have the user set variables > explicitly; as you said this happens rarely. It may be that setting > them automatically is not the right thing to do in many cases. If > they are updated automatically, each Fish instance has to know where > to get or send its updates, and configuring that may be too much > overhead for most use cases. Also, if the network connection breaks, > when it is re-established updates have to be merged and there may be > conflicts. > Perhaps I was not clear enough: I'm talking specifically about setups where ~/.config/fish is shared (via NFS or whatever) between several machines. So you won't have to configure anything. And if you home dir NFS dies, you shell is very quickly stuck anyway, so no conflicts ;-). In such a case sharing the universal variables is both "least surprise" and quite essential for uvars to be useful. I can deal at home with two and a half computers; but at work I routinely use a dozen machines with shared home.
Quick hack: I can deal manually by writing a couple of functions to push uvars from local fishd to a central file and pull them from the file to local fishd. Sort-of simple version control. In fact perhaps I'll really use version control to handle conflicts ;-). -- Beni Cherniavsky <[EMAIL PROTECTED]> (I read email only on weekends) ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Fish-users mailing list Fish-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fish-users