On Sun, 22.05.11 18:55, Randy Kramer ([email protected]) wrote: > > (Still cc'ing the list.) > > Lennart, > > Thanks for your response! That prompts some additional questions or > comments: > > * So, if I want to separate what I call my real user data (which in > terms of the definitions in GLib, would be the files that I'd put in > the "members" of the GUserDirectory enum), from all the other garbage > that ends up in $HOME, the way to do that would be to have all the > directories like G_USER_DIRECTORY_DESKTOP, > G_USER_DIRECTORY_DOCUMENTS, ... in $HOME, and then set > things like $XDG_CONFIG_HOME, $XDG_CACHE_HOME, $XDG_RUNTIME_DIR, and > $XDG_DATA_HOME to locations not in $HOME?
You never get to set XDG_RUNTIME_DIR on your own. That should be done by your OS. But sure, you can set those env vars as you suggest otherwise. > * At least partially at my instigation, it is intended (there is a > bug filed) that FHS 3.0 will include a reference to the XDG Base > Directory Specification in the /home section. I'm not sure that is > sufficient anymore. I'm thinking that there should be some reference > to the directories in the GUserDirectory enum. What do you think? Sure, why not. > * What does XDG stand for? (Has it changed since its inception?) I think it stands for "XDG Desktop Group" now. Recursive and stuff. Consider it synomymous to fdo. > * Comment--now I see from that GLib document on Miscellaneous Utility > Functions that it really is the intention that things like icons > customized for a particular user really are intended to go in > $XDG_DATA_HOME. (Somehow, looking at the XDG Base Directory > Specification, and not being aware of the GLib: Miscellaneous Utility > Functions document nor the wiki document on Software/xdg-user-dirs, I > expected that my "real user" data would go in $XDG_DATA_HOME (I mean > things like documents, photos, ...). Icons is app data, not user data. Hence xdg-user-dirs is not the right place for it. The fact that xdg-user-dirs is not specified in an XML docbook but in a wiki page has no influence on the validity of it. Consider it having the same relevance as xdg-basedir. Lennart -- Lennart Poettering - Red Hat, Inc. _______________________________________________ fhs-discuss mailing list [email protected] https://lists.linux-foundation.org/mailman/listinfo/fhs-discuss
