[Scummvm-devel] Fedora ScummVM package (Was: Merging TsAGE engine into trunk)

[Willem Jan Palenstijn]

On Fri, Apr 15, 2011 at 01:47:34PM +0200, Max Horn wrote:
> Am 15.04.2011 um 12:37 schrieb Willem Jan Palenstijn:
> > On Fri, Apr 15, 2011 at 11:31:27AM +0200, Max Horn wrote:
> >> Am 08.04.2011 um 23:03 schrieb Max Horn:
> >>> Could somebody who actually uses Fedora or at least knows a bit about it
> >>> (e.g. how one reports bugs properly :) file a report on this (and maybe
> >>> send the URL, so that we can track it / chime in if necessary)?
> >> 
> >> Is there anybody here who could please do this?
> >> 
> >> <rant>This time I swear I won't cave in and just do it myself, as so
> >> often. I mean, I don't use Fedora, not even Linux. If nobody else here is
> >> sufficiently bothered by this, then why should *I* care?</rant>
> > 
> > I've already contacted the Fedora maintainer just before our last release,
> 
> Thanks!
> 
> > but couldn't convince him to drop the enable-all-engines flag. (But I also
> > haven't pressed the issue.)
> > 
> > The reply was:
> > "
> > Well if there is at least one game playable in an engine its worth enabling.
> > It's really convenient for packaging as I do not have to --enable-sci
> > --enable-sci32 --enable-some-new-engine.
> > Of course I could use a set of --enable but I would rather enable
> > unsupported games/engine than disabling playable games.
> > "
> 
> OK, well I guess we should first try to discuss this some more with him and
> explain our reasoning. In particular, I'd ask him how he would feel if we
> took something he made, modified it in a way against his explicit wishes, and
> then distributed it without any hint to end users that such undesired changes
> were made? At the very least, they should call their package scummvm-testing
> or something like that then!
> 
> But hopefully he'll understand. If not, we'll have various options, including
> trying to appeal to some higher level institution, making our configure
> script stricter,  etc.

One thing which I tried without much success, was splitting up the scummvm
package into "scummvm" (containing the stable engines) and
"scummvm-experimental" (containing the unstable engines as plugins). This works
fine for the engines which are entirely disabled by default, but not so well
for disabled subengines.

If external plugins (with more subengines enabled) would override internal
"plugins", this might work, though.


-Willem Jan