Re: Checklist for packaging games
2008/4/20, Ivan Vucica <ivucica@gmail.com>:
> That will definitely help newbies like me. Thank you!
>
> I'd ask a question with regard to these points:
>
> * Game loads game data independent of the arch endianness and word length
> * Game loads and saves user data independent of the arch endianness
> and word length
>
> 1. Does this mean the packager is supposed to fix upstream's code in
> portions of loading data, in such a manner which clearly should be
> sent to upstream and not just applied in one distro?
Either that or convince upstream to fix it. I don't believe in
universal solutions that can be applied blindly to every problem, so I
think in terms of the possible solutions for each case. The best
solution would be to inform upstream of this problem and get them to
fix it. If they cannot, and you can develop patches for fixing it, you
should send them upstream so that they can apply it.
If upstream is unsensitive to this problem (like, hey, I code my game
in Windows and it works perfectly here, so I don't mind a bit about
your problem) then probably the best would be to respect the formats
with the endianness and word lenght they have i upstream versions, but
fix it in the rest so that it works (we did that in hex-a-hop for
example, as far as I can remember).
Some games are not maintained upstream any more, or upstream is not
very responsive, and thus Debian acts effectively as upstream. In this
case we have to take care of the problems and fix them ourselves.
Anyway, all of this has to be done carefully in case we want to keep
compatibility with data sets coming from different architectures, even
more if older versions of the game are already in the archive. In that
case the data stored by previous versions of the game must still be
loadable by newer versions, users shall not lose their data.
> 2. Is it absolutely necessary that e.g. savegames are shareable
> between architectures?
I don't know if it's a must. At least it's important to have it in the
checklist so that if you're ignoring it, you're doing it on purpose
and not due to not knowing about it. For the packages in which I'm
most involved, it's an important thing, so I take care about it. I
don't know how serious that bug could be considered. I guess it might
depend on the case.
Any thoughts about this?
Greetings,
Miry
Reply to: