2009
10.19

libtcod 1.5 postponed

While trying to work on my new prototype with libtcod 1.5, I realize that I have gone too far in the library restructuring, forgetting the only development rule I shouldn’t forget : always advance with small steps and always have a clean, compilable, working project at the end of each small step. The 1.5 branch is now full of half-finished, completely bugged stuff and requires a lot of work to go back to the light side of the force. In fact, much more than I can afford. Going further in that direction would put the whole project at risk.

That’s why I decided to keep working on the 1.4 branch, gently bringing 1.5 hot stuff back to it. For example, I’ll backport the subcell stuff and maybe colored tiles support in 1.4.3. The big changes like opengl support or alpha channel may come later. The good side is that people already using libtcod for their project won’t be stuck with the 1.5 huge API break.

The other good news (at least for me) is that I can go back to code my next prototype right now. This one should be much closer to the final game than the previous ones. In fact, it could even be a real game with a smaller range than The Chronicles of Doryen, but I still expect a few hours of gameplay to finish it. It relies on pyromancer code base. Here is a very early screenshot. Might remind you some old screens from TCOD…

thecave-sc0

5 comments so far

Add Your Comment
  1. Sounds good. Some of the stuff in 1.5 scared me (as the libtcod-net maintainer).

  2. For what it’s worth Jice, I’m glad. I was thinking that the sheer amount of changes at once was quite extreme. And if it means we get to see more progress on TCOD proper, so much the better!

  3. Hehe thanks for you support. Too bad I wasted so much time on the 1.5 (especially since I perfectly know leaving the micro-step method is bad, but apparently you can’t throw away bad habits that easily…). Fortunately, I realized my mistake before going into the wall…

  4. Wasted? Jice, you introduced or half-introduced many cool features. Some were a bit shaky, like the OpenGL support (remember the Linux screenshot I sent you, black and white and rotated by 180ยบ? That was a fun bug :P), but reintroducing them using ye olde small steppe method might bear the expected results. Plus, you know what obstacles you can expect.

    Oh, and maybe I won’t get lost in the 1.5 code base and contribute something useful again :D. Probably not this month, but I’m slowly planning to get back to coding, especially since the UR engine rewrite idea is getting mature in my head…

    Anyways, I’m sure getting back to the stable 1.4.2 code base can only result in positive things.

  5. Yeah, you’re probably right. The time spent on 1.5 is not completely wasted, it’s just a kind of expensive brainstorming ๐Ÿ˜‰
    Besides, the 1.4.3 and 1.4.4 releases will probably come much faster than the previous ones since I only have to backport stuff from the trunk…