Merge with BULL?

Would you like to see BBAI come with BULL?

  • Yes, without BULL I might not use BBAI

    Votes: 18 23.4%
  • Yes, without BBAI I might not use BULL

    Votes: 23 29.9%
  • Yeah, but periodic merges would be fine

    Votes: 21 27.3%
  • Whatever, I'm not picky

    Votes: 7 9.1%
  • No! I need BBAI pure

    Votes: 8 10.4%

  • Total voters
    77
Oh: on the periodic merging of code branches: As a programmer by trade, this only works if you merge frequently enough. Typically a merger every couple of days is painless. Let a couple weeks go by before a merger and it becomes pretty painful. Let several months go by and forget about it.

Correct me if I'm wrong EF. But I think the point of your suggested system was so that BBAI and BULL could share an SVN, and the devs of the two projects (mainly EF and jdog) could check in updates to a single project as new code was implemented.
 
Typically a merger every couple of days is painless. Let a couple weeks go by before a merger and it becomes pretty painful. Let several months go by and forget about it.

In general I agree with this statement, but we have a fairly disjoint set of changes here. Things I change shouldn't affect things jdog changes and vice versa. The downside to frequent merging is frequent merging. ;) As long as decreasing the frequency doesn't increase the testing required, it will work out.

Correct me if I'm wrong EF. But I think the point of your suggested system was so that BBAI and BULL could share an SVN, and the devs of the two projects (mainly EF and jdog) could check in updates to a single project as new code was implemented.

That's certainly a possibility we could discuss, but I was more suggesting that a general method of designing SDK mods. We already have BAI and the UP being included in other mods, and I suspect the same will happen with BULL. If we all make our changes using #ifdefs, it should be easier for others to maintain merges of several mods and produce multiple versions with the various bits turned on/off as desired.

Instead of merging BULL and BAI, we could keep them separate, convert them both to use #ifdefs, and then collectively create a single merge that contains a bunch of these community mods all rolled into one SDK. We'd develop our own mods individually for easier testing, but we'd produce multiple versions from the combined SDK. This might be a lot of work, but I think we could streamline the process, combine forces, and save work instead of having each player individually merge their own desired collection.

Imagine having a monthly release of the giant merged SDK that has updates from whichever mods had their own updates that month. Of course, the more mods that get included, the more built DLL versions need to be compiled. Anyone have access to a compile farm?
 
Yes, all of the hassles of merging involve the same file changing on both branches.
The headaches occur when the same lines on the same file changed differently.
The forget about it is when someone has auto reformatted the files / renamed a file / shared public method.
 
Yeah, BBAI includes the unoficial patch. I'm not sure if it includes Alexman's amphibious assault fix, but I assume it does, if it doesn't jdog must have missed it, and it's supposed to be/will be in there.

Yeah, I caught Alexman's amphibious assault fix, it's in there as well. As well as a few other things I've found which could be in the Official Patch ...
 
In general I agree with this statement, but we have a fairly disjoint set of changes here. Things I change shouldn't affect things jdog changes and vice versa. The downside to frequent merging is frequent merging. ;) As long as decreasing the frequency doesn't increase the testing required, it will work out.



That's certainly a possibility we could discuss, but I was more suggesting that a general method of designing SDK mods. We already have BAI and the UP being included in other mods, and I suspect the same will happen with BULL. If we all make our changes using #ifdefs, it should be easier for others to maintain merges of several mods and produce multiple versions with the various bits turned on/off as desired.

Instead of merging BULL and BAI, we could keep them separate, convert them both to use #ifdefs, and then collectively create a single merge that contains a bunch of these community mods all rolled into one SDK. We'd develop our own mods individually for easier testing, but we'd produce multiple versions from the combined SDK. This might be a lot of work, but I think we could streamline the process, combine forces, and save work instead of having each player individually merge their own desired collection.

Imagine having a monthly release of the giant merged SDK that has updates from whichever mods had their own updates that month. Of course, the more mods that get included, the more built DLL versions need to be compiled. Anyone have access to a compile farm?

Wow, this is right in line with the "next step" Johny and I have been talking about for the WoC Lite/Core/SDK/Modular/Thingy :)

I am just getting in to C++/Python where my previous programming is from many years ago in Fortran/Turbo Pascal/BASIC.

Once I feel more comfortable in the SDK then I hope to have a round-table of modders to get ideas just like yours EF. The power of modularity in Civ4 BtS has yet to be fully realized but is getting very close...even without Firaxis doing much "right".

The idea of modular loading has been out long enough and I see a number of people using WoC or doing some of their own modular things so it seems like we are close to bringing everybody together to make our modding easier for us and more enjoyable for the players.
 
So, what should we call it? BUBBA? :mischief:

:lol: EIEIO? I've always wanted to think up a name that would work with that acronym.

Are you going to support 3.17 or move to 3.19 completely? I've already decided to go to 3.19. Most of the reasons people weren't using the UP 0.21 were because it wasn't "official", but now that that argument is toast, I figure what the heck. Dealing with all the merges is going to be enough of a headache; I don't want to add in multiple SDK versions.
 
bull+bai+(new)UP+3.19?

share? (I could care less if beta code eats my game midway thru, I'm a great abandoner of games anyway)
 
This is great news, I appreciate the work from everyone :)
 
whats the ETA?
 
I have too many things stacked up in front of it that it would be pointless to pick a date. Probably less than a month from now is the best I guess I can make.
 
What about the idea of you and jdog sharing an SVN so that new updates check into each other and you guys develop a standardized core? Did this just not work out?
 
We're going to maintain a shared SVN for BULL + BBAI. As we make updates to our own SVNs, we'll occasionally check in the collected changes to this new SVN.
 
I wanted BTAI+BUG+CarterEarth, so I mashed them together. Hardest part was adding a few dummy civs to the scenario to make it work. Nothing is lost, since there were no file collisions. You maybe have to reinstall Blue Marble.

Extract into your mod folder.

+ Better AI 0.78c with 48 civs
+ BUG (latest nightly, 3.6? whatever)
+ CarterEarth32 scenario, my favourite earth map by far
 

Attachments

Top Bottom