On Sun, 03 Jan 2016 21:03:13 +0100, Alexandre Detiste <alexandre.detiste@gmail.com> wrote: > I have copied the implicit install_as: (manual.pdf) to the look_for: list > to make sure the .pdf doesn't get lost when repacking the CD game from > /usr/share/games/... (this can be usefull if the .desktop file were improved > or an icon were provided by GDP). > > files: > manual.pdf?cd: > - look_for: [Sam & Max Manual.pdf] > + look_for: [manual.pdf,Sam & Max Manual.pdf] Thanks! > Now, maybe this operation (always adding install_as to look_for) > should allways be done in the python code instead ? I think that would be better, it seems weird to have to include the target name (especially since it's useful only for repacking). > > I've been going through my old CDs, and I've come across a few games which > > differ from the variants supported in game-data-packager. Many of these > > CDs don't include a license file of any sort, and I'm wondering how to > > handle that when the current packaging expects one. For example, in The > > Dig (dig.yaml), the-dig-en-data and so on list "end user license > > agreement.txt" as a license file; since the CD doesn't include it, g-d-p > > goes and looks for it elsewhere, and fails if it can't find it. > > It fails because it can't find both the GOG & CD version at the same time > as requested in the yaml file... but you already found out (see your next > message) ;-) > > I would say from the reverted commit that you'd need alternatives like: > > video/font1.nut: > alternatives: > - video/font1.nut?gog > - video/font1.nut?cd > > And then font*.nut?* must be moved out of "shared assets:" group into > a new group that doesn't get included in the packages. > > I usually name it "default:", but there's no rule, > "misc files:" or "special files:" would be ok too. > > Then the alternative (video/font*.nut) would be listed into the package > files lists. Thanks, I'll give that a shot. > Archives, like the one provided by GOG.com can only provide real files, > not virtual ones (alternatives); so you'd need to list the "?gog" one > there. There's an assert() to catch this. That makes sense! Regards, Stephen
Attachment:
pgpLS4jdUMegH.pgp
Description: OpenPGP digital signature