Le vendredi 2 octobre 2015, 00:26:40 Simon McVittie a écrit : > Control: tags 800596 + pending > > On Thu, 01 Oct 2015 at 16:22:56 +0200, Alexandre Detiste wrote: > > Here's a regression in the 'steam' mode. > > I believe this is now fixed in git. I hadn't applied the necessary > separation between game data and packaging state in that mode.
Thanks > > > import'ing PackagingTask and doing instead: > > - for path in gamedata.iter_steam_paths((package,)): > > + for path in > > PackagingTask.iter_steam_paths(gamedata,(package,)): > > It is never correct to take an unbound method like > PackagingTask.iter_steam_paths and call it with a first (self) argument that > is of a different type (I think I fixed a couple of instances of that > pattern recently). GameData and PackagingTask are unrelated types. > > S I did more OO while working on this than at school; I should re-read my book. > The gog mode looks as though it should be OK (it doesn't start a task, > it only inspects game data), but please check. Yeah it's mostly a mockup of a TreeView now, users have to call GDP again for each game; but I find it already quite usefull, and it helped me to find a missing data tag for gobliiins game.