Checkmate50
Checkmate50
Would be fun to black-box fuzz test this thing, and is on my list of things to-do. Nothing fancy really, just a fun miniproject at some point.
Basically a amalgamation of #4 and #5 , this option would make things not use nearly as much stack and maybe actually terminate sometimes.
Another speedup / performance issue, but basically having an option for innter to be written to variables to avoid duplicate computation would be quite helpful for our poor stack. Very...
Let's allow big numbers! I want to use some clever binary representation for numbers instead of numerals to allow for actually reasonable integer sizes (already sorta done when interpreting literals...
How can you have a lambda-based language with lambdas? Riddle me this! Anyway, this will probably involve some syntax and scoping changes, so future problem, but the actual implementation in...
Very useful, will probably be implemented using HD / TL and pairs
Obviously necessary. Basically compile everything (including control flow and assignments) into a single line of code, probably using maps of some sort. Assignments are the tricky part here, need some...
Messages generated by the CUDA renderer fail to write to standard output. The writer fails on [line 58](https://github.com/shader-slang/slang/blob/master/tools/gfx/cuda/render-cuda.cpp#L58) with a read access violation (`this was 0x10`). Calling `StdWriters::initDefaultSingleton()` manually beforehand...
When a `__GPU_FOREACH` is executed, the resulting function is expected to have exactly three arguments: `renderer`, `gridDims`, and `buffer`. In real programs, this last argument could realistically be several arguments,...
Currently, the `-heterogeneous` flag attempts to write the bytecode blob of a program to to C++ (if possible). However, if compilation fails, the error produced is the internal error: `(0):...