I'm Back!
<i>This post was originally made by <b>danij</b> on the dengDevs blog. It was posted under the categories: Blog, Engine, Mac OS X, Platforms, Unix/Linux, Windows.</i>
As evidenced by my recent svn commits, my real life commitments have loosened their grip enough so as to allow time for deng again. Hopefully, things will stay this way for a while so that I might get some of the outstanding work for Doomsday 1.9.0-beta6 done.
My current plan is to continue working on the revised map loading process.
So far, things have been falling nicely into place and I've now got a local branch where I've internalized the core logic of glBSP 2.24 into Doomsday itself. The node building process now happens entirely on objects in Doomsday-local memory and with no need for temporary wad files in the working directory.
Although the map loading is not yet completely functional, I'm very happy with the way it is coming together.
The next step is to refactor the old finalizeMapData() as now that we have access to the map data structures constructed during the node building process; we no longer have to recreate them or waste time ensuring they are formed correctly.
After that, I'll be examining the frontlines between Doomsday and dpMapload (currently not involved in the process at all) to determine where the API should be.
Once this work is complete we can begin thinking about a caching stratergy. One idea could be to combine this with a new save system.
As evidenced by my recent svn commits, my real life commitments have loosened their grip enough so as to allow time for deng again. Hopefully, things will stay this way for a while so that I might get some of the outstanding work for Doomsday 1.9.0-beta6 done.
My current plan is to continue working on the revised map loading process.
So far, things have been falling nicely into place and I've now got a local branch where I've internalized the core logic of glBSP 2.24 into Doomsday itself. The node building process now happens entirely on objects in Doomsday-local memory and with no need for temporary wad files in the working directory.
Although the map loading is not yet completely functional, I'm very happy with the way it is coming together.
The next step is to refactor the old finalizeMapData() as now that we have access to the map data structures constructed during the node building process; we no longer have to recreate them or waste time ensuring they are formed correctly.
After that, I'll be examining the frontlines between Doomsday and dpMapload (currently not involved in the process at all) to determine where the API should be.
Once this work is complete we can begin thinking about a caching stratergy. One idea could be to combine this with a new save system.