[Xastir-Dev] Re: New Directory Structure

[Alan Crosswell]

Sounds good.

Curt Mills wrote:
> On Tue, 24 Jun 2003, Curt Mills, WE7U wrote:
> 
> 
>>On Mon, 23 Jun 2003, Alan Crosswell wrote:
>>
>>
>>>I have another suggestion for a directory change:  The places GPS maps are
>>>*written* needs to be configurable, and should probably be in ~/.xastir/ by
>>>default;  In a normal multi-user system, /usr/share/xastir/maps/GPS would not be
>>>writeable by mere users.  A knowledgable super-user could of course take
>>>shapefile maps from his .xastir/... directory and move them to
>>>/usr/share/xastir/maps as needed.
>>
>>I'm struggling with the best way to go about that.  I think I'd like
>>to modify the map chooser stuff so that it's a reasonably seamless
>>integration between local user maps and the system map directory,
>>including the GPS directory stuff.
>>
>>I don't want to get into changing too much code for the map chooser,
>>indexing, and map finding routines.
>>
>>Perhaps changing all of the routines so that they use the local user
>>map directories as the root, and then making a symlink from there to
>>the system map directories would be a good compromise.  Just
>>thought of that as I was composing this reply.  That might save a
>>lot of work.
> 
> 
> How about this:
> 
> 
> ~/.xastir/maps/system -> /usr/share/xastir/maps
> ~/.xastir/maps/<user>/
> ~/.xastir/maps/<user>/gps/
> 
> 
> In this case, Map Chooser would start at ~/.xastir/maps and set up
> to traverse symlinks, so you'd see "system", and "<user>" as the top
> two directories in Map Chooser.  Perhaps "public" and "private"
> might be better labels?  Whatever gets the job done.
> 
> Xastir would check for the directories/symlink, creating each if not
> present.
> 
> 
> ~/.xastir/maps/<user>/      Individual user maps
> ~/.xastir/maps/<user>/gps/  GPS downloads and APRS trail saves
> /usr/share/xastir/maps      System-wide maps
> 
> 
> "gps" could also be broken out into "gps" and "aprs-trails".
> 
> I'll have to check whether the code can handle symlinks as it's
> searching for/displaying/indexing maps.  I think it can already
> though, in which case the above scheme would be very easy to
> implement.
> 
> Thoughts?
>