Not the only reason. It seems COG does something else with their files when they put them on the stores, but if you want to compile a game so you can host anywhere other than Dashingdon’s (or compile it to host on Dashingdons rather than adding individual txt files), they need to be on the scene list. I also ran into a problem just the other day when I realised I’d left off a file on the startup menu so it didn’t run via the local testers packaged with the folder. (Again sounds like less of a problem if you are using CSIDE, but I still use the packaged tools in the folder myself at times, particularly if I’m fiddling around with the source code to include in the games for example.) In the end I guess it depends on what you want to do with the game, but it doesn’t hurt in any way to have the text files listed on the scene_list and in some cases might be necessary.
Hmmm. Does that affect randomtest and quicktest? I never thought about that. I am using the IDE for such a long time that I didn’t know about these differences…
I think CJW said it didn’t, it’d only be if you were using the internally packaged testers and run files I’d think.
I’ve never had to list all files to test locally using any build of choicescript from 2016 to current. I’ve never used CSIDE, either. I make my games entirely using Notepad++ and the latest public choicescript build. I’m not sure why you’d be experiencing this behavior, but I assure you I’ve never encountered that issue. 
Actually, I disagree.
You thoroughly test your games, which means I don’t get on you about this.
But in general, all scene files should be in the *scene_list, so that Quicktest will review it. If it’s not in the list, QT won’t review, and there could be hidden bugs.
And the problem with putting *bug at the top of a scene file is that QT will crash on that.
That’s fair.
I didn’t know that QT doesn’t check non-listed scenes. I know Randomtest does, so I assumed QT did as well.
I thought QT did all scenes in the folder too!
So would it be a case of using *ending rather than *finish at the end of the final played scene, to avoid falling into non-chronological scene files? Or would that also prevent those scenes getting tested properly?