Re: Translations of folder names - two proposals
- From: Havoc Pennington <hp redhat com>
- To: Alan Cox <alan lxorguk ukuu org uk>
- Cc: Alexander Larsson <alexl redhat com>, "desktop-devel-list gnome org" <desktop-devel-list gnome org>, Jamie McCracken <jamiemcc blueyonder co uk>
- Subject: Re: Translations of folder names - two proposals
- Date: Sat, 11 Dec 2004 15:01:09 -0500
On Sat, 2004-12-11 at 18:41 +0000, Alan Cox wrote:
> > $HOME/Desktop does still inherit the $HOME problems, but it does not add
> > a $DESKTOP problem. For example, say we add a gconf feature as we've
> > discussed to solve the second problem above:
> > gconf_client_set_filename() where /home/hp is substituted with ~
> > gconf_client_get_filename() where the ~ is changed to /home/hp
>
> Does gconf not support recursive late resolution of keys containing
> other gconf keys ? Perhaps that is a - long term - solution to such
> things.
That could work because it lets you introduce new "variables" in the
config value that would still be known to old programs. That's the
advantage over some hardcoded list of substitutions for the config
value, is that you could avoid
gconf_client_set_filename_with_newer_substitutions().
However it introduces one big world of pain: change notification. If a
key changes, you need to know all keys that reference that key and also
notify that those keys have changed. This is why I didn't implement
"symlinks" in gconf, because as far as I could tell I would have to load
the entire database, and create a mapping from symlink targets to
symlink names, and when doing a change notification use this mapping to
propagate additional change notifications. A couple other issues come up
as well such as cycles.
Havoc
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]