Re: [AD] 5.0.0 final release plan |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On January 2, 2011, Matthew Leverton wrote:
> I haven't really read all of this thoroughly, so apologies if I'm
> repeating something. But why not just have something like:
>
> ALLEGRO_PROGRAM_DATA_PATH
>
> If executed from a bundle, it points to the Resource folder (if it
> exists). If not in a bundle, it's the same as ALLEGRO_PROGRAM_PATH.
>
> Because to me, it seems like ALLEGRO_SYSTEM_DATA_PATH is something
> like c:\users\all users\appdata\game\, which makes no sense in terms
> of locating data that is bundled with a program. And
> ALLEGRO_USER_DATA_PATH is something like
> c:\users\matthew\appdata\game\, which again, makes no sense.
SYSTEM_DATA_PATH is/was intended to be where data that comes with a game is
installed to, on installation. I'm not entirely sure what that maps to these
days in vista/w7. It used to map to C:\Program Files\SomeGame.
I assumed it would still map to a similar place, one that a user wouldn't be
able to modify, much like /usr/share on unix.
> To me, both of those (SYSTEM_DATA and USER_DATA) are meant for saving
> content generated by the program. For example, if I want to save a
> high score, I might save it into SYSTEM_DATA (if shared among all
> users) or USER_DATA (if just for that user). But I would never use
> those locations to find data that is bundled with the program.
Yeah, SYSTEM_DATA was intended to be read only. At the very least, you'd have
to expect it to be read only at least on one of the supported platforms, so
you might as well treat it that way for all platforms.
> Now I know Linux is slightly different in terms of some installation
> spreading data away from the executable, but to me, that's something
> the program should have to explicitly look for (via SYSTEM/USER_DATA).
>
> In any program I would write, I would simply assume the data is always
> bundled with my program. So I would use something like
> ALLEGRO_APP_DATA_PATH for that. If the user moves it to some other
> location, it won't work, and that's his fault.
>
> --
> Matthew Leverton
>
I don't really care one way or the other what each returns, so long as its
sane and consistent for each platform.
--
Thomas Fjellstrom
tfjellstrom@xxxxxxxxxx