The current state is a total mess, several things don't work right and it seems doubtful that this can really be sorted out on short notice.
This is a huge problem because even useful additions are getting lost in this problematic branch.
With that in mind I'd like to ask to set the master branch to a stable commit and then carefully add the useful things that were done recently, without that 'level refactoring' that is going on, for whatever reason I cannot see.
Reset 'master' branch to a stable commit
Moderator: GZDoom Developers
Re: Reset 'master' branch to a stable commit
This cannot be done as requested. We have the 3.7fixed branch, and you can compile from that if you really need something recent and stable.
But as it stands, we actually need *more* people to test the master branch and report bugs on it. Understand that refactors like that are being done for a reason - I get it, it's inconvenient, and sometimes I find myself switching to my 3.7.x builds for normal play because of it. But in the end - the features they might enable in the long run will be well worth the inconvenience.
But as it stands, we actually need *more* people to test the master branch and report bugs on it. Understand that refactors like that are being done for a reason - I get it, it's inconvenient, and sometimes I find myself switching to my 3.7.x builds for normal play because of it. But in the end - the features they might enable in the long run will be well worth the inconvenience.
- Graf Zahl
- Lead GZDoom+Raze Developer
- Posts: 49056
- Joined: Sat Jul 19, 2003 10:19 am
- Location: Germany
Re: Reset 'master' branch to a stable commit
I'd say once I find time to test yesterday's commit about moving the thinker lists into the level we really need to do a 3.8 alpha so that the code gets wider exposure.
I'd wait with further feature development until everything is more stable - but that thinker list move is the crucial piece of the whole puzzle, without that, testing is pointless.
Once that is in place we may also consider a proper compatibility layer for the whole thing, but that needs some data first, i.e. loading and running mods, and the more people help with that, the better.
I'd wait with further feature development until everything is more stable - but that thinker list move is the crucial piece of the whole puzzle, without that, testing is pointless.
Once that is in place we may also consider a proper compatibility layer for the whole thing, but that needs some data first, i.e. loading and running mods, and the more people help with that, the better.