HOME | DD

Description

Once software development reaches a certain stage, and it occurs fairly early on, it is a good idea to have a set of standard tests that you run after each program change. What I am doing is developing python scripts for blender 2.79b to import and export *.msh files from "The Movies" a computer game by Lionhead (RIP). We currently have scripts written for blender 2.49b and the goal is to modify or rewrite as necessary these scripts to be able to form the same game asset modification tasks in blender 2.79 that are now performed in blender 2.49b. The reason this is not a trivial task is the drastic change in the blender python application programmer's interface (API) between version 2.4 and 2.5

We now have a script which will import msh files, although it is somewhat limited in capability because certain features were omitted. The focus of my present effort is to produce an export script to go with the import script.

The above chart shows the tests I am running to verify capabilities and test software changes. Each line of the chart represents a different series of tests. On the first line, we use the old script to import a mesh file (the wooden mannikin), save it as a blend, then re-export the mesh to the game. We then run the game to see whether the exported mesh will work in the game. The result of this test is shown: the test was successful, as expected.

The second line is a closure test. In the closure test the export mesh file is re-imported using blender 2.49b just to see if can be loaded closure test). This was successful as well, not surprisingly.

The third line is the reason for this report. The mesh export script for blender 2.79 has reached a point where I was able to open the blend file containing the manikin, then export it to the game folder and run the game. This is a major step forward. The fourth test, on the next line, is the closure test. My current version of the import script for blender 2.79 was not able to correctly re-import the mesh file from the game folder, although it will import the original mesh file (woodenman) extracted from the PAK files. Once I am able to complete the closure test for export script, I can move on to the fifth row, which is testing of import and export scripts in combination. Hopefully, by that point, the export script will have been proven reliable and analysis and development can focus on the import script.

The empty circles are empty because at this point, I am not able to perform these tests. The export script fails when I attempt to use it in combination with the import script.

Once all the circles can be filled in the world "success" and the process has been repeated for a bunch of different msh files, it should be possible to lay blender 2.49 aside at last.

Description

Once software development reaches a certain stage, and it occurs fairly early on, it is a good idea to have a set of standard tests that you run after each program change. What I am doing is developing python scripts for blender 2.79b to import and export *.msh files from "The Movies" a computer game by Lionhead (RIP). We currently have scripts written for blender 2.49b and the goal is to modify or rewrite as necessary these scripts to be able to form the same game asset modification tasks in blender 2.79 that are now performed in blender 2.49b. The reason this is not a trivial task is the drastic change in the blender python application programmer's interface (API) between version 2.4 and 2.5

We now have a script which will import msh files, although it is somewhat limited in capability because certain features were omitted. The focus of my present effort is to produce an export script to go with the import script.

The above chart shows the tests I am running to verify capabilities and test software changes. Each line of the chart represents a different series of tests. On the first line, we use the old script to import a mesh file (the wooden mannikin), save it as a blend, then re-export the mesh to the game. We then run the game to see whether the exported mesh will work in the game. The result of this test is shown: the test was successful, as expected.

The second line is a closure test. In the closure test the export mesh file is re-imported using blender 2.49b just to see if can be loaded closure test). This was successful as well, not surprisingly.

The third line is the reason for this report. The mesh export script for blender 2.79 has reached a point where I was able to open the blend file containing the manikin, then export it to the game folder and run the game. This is a major step forward. The fourth test, on the next line, is the closure test. My current version of the import script for blender 2.79 was not able to correctly re-import the mesh file from the game folder, although it will import the original mesh file (woodenman) extracted from the PAK files. Once I am able to complete the closure test for export script, I can move on to the fifth row, which is testing of import and export scripts in combination. Hopefully, by that point, the export script will have been proven reliable and analysis and development can focus on the import script.

The empty circles are empty because at this point, I am not able to perform these tests. The export script fails when I attempt to use it in combination with the import script.

Once all the circles can be filled in the world "success" and the process has been repeated for a bunch of different msh files, it should be possible to lay blender 2.49 aside at last.