Path of the Shell did not like my fancy metaprogramming tricks for
defining an AgeSDL Python class that contained characters that are
illegal in Python identifiers. So, now, we revert to just using a
standard class declaration.
That means that we need to strip out any illegal identifiers from the
age name first. A legal Python 2.x identifier is constrained to the
ASCII alphanumeric characters and the underscore with the stipulation
that the first character cannot be a number. To illustrate this to the
artist, we alert the age name property field if an illegal character is
found in the age name. We also alert on the underscore, which is now
used as a very very special replacement character. In the case of an
illegal character, an error message is shown in the UI with the correct
AgeSDL name.
Of course, I hope no one really uses those illegal characters and this
is just more fulmination on my part...
There are some cases where errors, while bad, are not the end of the
world. I'm thinking namely about compyling the age python. The age still
exports just fine, but the ancillary data is flawed. This new system
collects nonfatal errors until the export is done, then raises them all
at once.
Implements the boilerplate code for compiling Python code in arbitrary
python versions and packing the marshalled data into Cyan's Python.pak
format. Since this is a lot of bp, a separate operator has been added
to both test the resulting mayhem and provide age creators an easy way
to export only their needed Python.
The only python that is packed currently is the age sdl hook file, if
any. In order for that part to happen, Python File nodes need to be
upgraded from having a string path to actually using the new text_id
field.
Age output files are now handled in all aspects by a singleton manager.
This allows us to track all generated files and external dependency
files and ensure they are correctly copied over to the target game... Or
not, in the case of an age/prp export from the File > Export menu.
Currently only SFX files are handled as an external dependency. TODO are
python and SDL files.
Further, because we have an output file manager, we can bundle all the
files into a zip archive for releasing the age in one step. Wow such
amazing. ^_^
If the texture cache doesn't return images in exactly the order or way
that libHSPlasma is expecting, it raises a RuntimeError. We can detect
that we used a cached image and regenerate the data in that case...
instead of just outright failing with "image data size mismatch".
This operator takes a file as an argument and builds a cubemap from it.
Valid options are to supply the output from Plasma's
Graphics.Renderer.GrabCubeMap console command. The operator will find
the other five files and generate a cubemap with the faces saved by
Plasma. Otherwise, any arbitrary image can be supplied. If the filenames
do not fit the expected format, any missing faces will be replaced by
the face specified in the file selector. This will generally result in a
cubemap with six identical faces.
Previously, we allowed OpenGL to generate all of the mip levels for us
in a mipmap. This was pretty doggone fast and worked reasonably well.
However, with cube maps, we will need to use images that are not always
backed in Blender... this is because Blender stores cube maps as one
single image instead of one image per face. So, we need to be able to
generate those mip levels, preferably without touching Blender's
`Image.pixels`, which is slower than Christmas...
Also of note... `Image.gl_load()` will actually scale the iamge to a POT
when Blender is using OpenGL ES... but not on other platforms. So, now,
we just ask Blender to load the image and deal with the POT-izing later.
The con here is that the pure python implementation of the image scaling
function is SLOOOOOOOW. We're talking ~40 seconds to process a 1024x1024
mipmap. No one should be using the reference implementation, however,
and the C++ implementation shows no noticable slowdown over the OpenGL
code.
Whew.
Cube maps in Plasma are six separate images; however, Blender expects to
be supplied with one image that encompasses all of the cube faces.
Therefore, our texture cache needs to be able to support many images for
one request.
This provides the artist the ability to disable exporting layers that
would be DynamicCamMap (fast) in MOUL but DynamicEnvMap (slow) in PotS
in one fell swoop. Please note that disabling env map exporting will NOT
prevent waveset environment maps from exporting.
The lightmap modifier is now a more general "Bake Lighting" modifier
that can control how vertex colors are baked as well. The default vertex
color baking behavior is unaffected. It is now identical to adding a
Bake Lighting modifier, changing it to bake vertex colors without
specifying a pass.
Unfortunately, packed images don't appear to store any kind of modify
time attribute. In the case of those, we're just idly hoping that we can
find one on disk.
What was previously marked as transition cuts would really only cut goal
and POA tracking. Transition cutting requires the addition of a
transition ovveride. The engine supports cutting transitions on
plCameraMsg receipt by the virtual camera. The option looked sloppy on
regions (and should probably be controlled in the transitions), but it
has been added to camera message nodes for more control.
Given that the Plasma World (read: age) is defined per scene, it seems
appropriate to limit the export to the current scene. Other scenes could
potentially have other age definitions...
Arbitrary data from a material's layer flags were able to creep into a
DSpan's criteria, causing geometry that should go into the same span to
spawn a new DSpan with an identical name. Only one of the DSpan objects
would be exported, and Uru would either show crazy stuff or crash as a
result.
On the surface, this allows lightmaps to be exported as PNGs. However,
it also includes a light refactor of image exporting to facilitate
requesting export formats and the targeting of plImageLibMod. This
should better enable #96.
Due to our disabling of texture baking when lightmapping, Blender would
not be able to cast light through transparent materials. Now, we revert
back to something resembling the manual method of lightmapping from the
bad old days of PyPRP.
Do note that because we attempt to batch object baking as much as
possible, I have disabled baking lightmaps to transparent objects.
This is because the transparent objects would need its textures disabled
for it to be baked to, but with the textures disabled, the renderer
would be unable to correctly determine the object's transparency. D'oh!
It would be possible, I suppose, to move them into their own batch, but
that makes the state machine a bit more complex. It would be better to
wait for someone to need that functionality, I think.
