Try mmapping the whole file, so we never have to worry about refilling the input buffer.
- Don't need to worry about copying values into internal storage that way either, can copy them directly into user-provided buffers as part of the extraction step instead.
Provide a way to handle files where faces need triangulation, but we haven't read the vertex data yet:
- Record the byte offset in the file of each element, as we reach it.
- Allow jumping directly to the start of any element with a known offset.
- Calculate offsets up to the first non-fixed-size element at startup.
Expand the docs.
Set up continuous integration
- Github Actions?
- Travis?
Automated tests to ensure we can parse all file and data types correctly.
Option to generate graphs of the miniply-perf results as SVG files?
Send a pull request adding miniply to the ply_io_benchmark suite.