Re: [PATCH] gnu: Add openttd

[Leo Famulari]

On Tue, Apr 12, 2016 at 04:55:20PM +0200, Albin wrote:
> Thank you Leo for taking the time to inspect and install the package.

Thanks for taking the time to work on it :)

> Den 2016-04-12 kl. 03:09, skrev Leo Famulari:
> > On Tue, Apr 12, 2016 at 12:41:58AM +0200, Albin wrote:
> >> +                 ;; GNU's not Unix so let's modify the exit dialog.
> >> +                 (substitute* (find-files "src/lang/" "\\.txt")
> >> +                   ((":Unix") ":GNU"))))))
> > 
> > I don't think this is necessary.
> 
> Maybe not.  I just wanted to exercise my freedom a little.  The above
> snippet lets you see this dialog box:
> 
> "Are you sure you want to exit OpenTTD and return to GNU" ...instead of
> "...Unix", which I thought would make the Guix users happy.

It did make me smile :) I appreciate the motivation behind this change
very much.

> It works for 54 language versions, but not Arabic, Esperanto, Hebrew and
> Urdu where "Unix" is called something else.
> 
> I'm not going to insist on keeping it, but remember that the "Words to
> avoid" article (https://www.gnu.org/philosophy/words-to-avoid.html)
> states: "To call the whole system 'Linux' is both unfair and confusing."
> Isn't it likewise wrong to call "GNU" "Unix"?

I'm open to a wider discussion, but I don't think we should patch
upstream code unless it's to fix bugs.

Plus, I personally think it should say "return to Debian", since that's
what my workstation is running ;)

> > I see this option in `./configure --help`:
> > 
> > --with-liblzo2[=liblzo2.a]     enables liblzo2 support
> 
> Nice!
> 
> > Does it work to provide the path to that option? Our lzo package does
> > build a static library, if that is what OpenTTD requires.
> 
> Yes, I could make it work by first installing 'lzo' and then add
> `--with-liblzo2=/gnu/store/[...]/liblzo2.a` to the package definition.
> What is the proper way to do this?

There's a clear example in the package definition of 'apl'. Basically,
make 'lzo' an input, and then build a string like this...:

(string-append "--with-liblzo2=
               (assoc-ref %build-inputs "lzo") "/lib/liblzo2.a")

... and pass that as an argument to ./configure. You typically get a
useful error message from ./configure if there's a mistake.

> >> +    ;; The software contains an in-game downloader from which the user
> >> +    ;; may find non-functional data licensed under different terms.
> > 
> > I don't know the significance of this. Hopefully somebody else will
> > comment on this subject.
> > 
> > Upon starting the program, I was prompted to download something related
> > to graphics. I agreed, and the game seemed to download OpenGFX [0] and
> > launch correctly, although it also warned me about needed to download
> > some sound files.
> 
> Yes, it's not ideal that you are immediately prompted to download and
> install something like this.  As you said, the package that gets
> installed when you agree is OpenGFX (8 MiB).  It has the same license as
> the game itself, i.e. GPLv2, just like the other available graphics sets
> "NightGFX" (3.4 MiB) and "zBase" (273 MiB).
> 
> The music files that you can download are:
> 
> * Modern Motion (music): CC-BY-NC-ND 3.0
> * OpenMSX: GPLv2
> * Scott Joplin Anthology: public domain
> * Traditional Winter Holiday Music: public domain
> 
> For sound effects only one package is currently available:
> 
> * OpenSFX: CC Sampling Plus 1.0
> 
> 
> I can understand that packages that restrict commercial redistribution
> may not be included in Guix, but the FSDG doesn't discuss the particular
> case of packages that *recommend* non-free, non-functional data where
> only non-commercial redistribution is permitted.
> 
> How does other free distros deal with this?  The Arch project
> distributes openttd, openttd-opengfx and openttd-opensfx as separate
> packages (where the package manager recommends the latter two upon
> installation of openttd).  Parabola does the same, but removes
> openttd-opensfx from their repository.  You can still download the sound
> effects afterwards however.
> 
> Do you think we should follow Parabola in this case or take a different
> approach?

I don't know the details of how we are supposed to handle these issues.
I'll let more knowledgeable people have this discussion.