Print name of locals if possible in disassembler |
|
More...
|
almost 4 years ago
|
Fix merge issue, use helper to compute jump offset |
|
More...
|
almost 4 years ago
|
Don't optimize field reads in blocks as getters |
|
More...
|
almost 4 years ago
|
Added support for the new bytecodes for the trivial method optimizations, and fix inlining of trivial methods |
|
More...
|
almost 4 years ago
|
Added negative tests for trivial method optimizations, and tests for bytecode inlining of trivial methods |
|
More...
|
almost 4 years ago
|
Add support for trivial blocks, and make sure only setters can have 2 statements while being trivial |
|
More...
|
almost 4 years ago
|
Specialize SObject.toString() for use in tests, add check for `nil` |
|
More...
|
almost 4 years ago
|
Fix exception preventing parse errors to show for strings not containing new lines |
|
More...
|
almost 4 years ago
|
Added tests for trivial method optimizations |
|
More...
|
almost 4 years ago
|
Added tests for return_field_? bytecodes |
|
More...
|
almost 4 years ago
|
Merge PR #80: Make symbol table global, to avoid indirection to universe |
|
More...
|
almost 4 years ago
|
Make symbol table global, to avoid indirection to universe |
|
More...
|
almost 4 years ago
|
Merge PR #79: Remove second bytecode pass for stack hight and use symbols for checks |
|
More...
|
almost 4 years ago
|
Remove second pass over bytecodes to determine stack hight |
|
More...
|
almost 4 years ago
|
Use array to encode bytecode stack effects |
|
More...
|
almost 4 years ago
|
Use symbols to avoid run-time string comparisons for true/false/nil |
|
More...
|
almost 4 years ago
|
Add support for true/false/nil LiteralNode creation |
|
More...
|
almost 4 years ago
|
Simplify some bytecodes, and always allocated the quickened array |
|
More...
|
almost 4 years ago
|
Set the second byte of the jump offset 0, always, just to be sure |
|
More...
|
almost 4 years ago
|
Avoid issue with not adopted nodes in disassembler during inlining |
|
More...
|
almost 4 years ago
|
Fix jump offset reading not taking into account signed byte semantics |
|
More...
|
almost 4 years ago
|
Fix assertion to allow all jump bytecodes |
|
More...
|
almost 4 years ago
|
Added dump(byte[]) to test setup |
|
More...
|
almost 4 years ago
|
Fix backwards jumps as in PySOM, using a list of inlined loops/jumps |
|
More...
|
almost 4 years ago
|
Fix offset calculation in Disassembler |
|
More...
|
almost 4 years ago
|
Adjust some tests to TruffleSOM bytecode set (it differs around frame operations) |
|
More...
|
almost 4 years ago
|
Make emitPUSHGLOBAL produce a push_constant for known constants, to have same semantics as in AST interpreter |
|
More...
|
almost 4 years ago
|
Make sure shortest possible PUSH_FIELD bytecode is used after inlining |
|
More...
|
almost 4 years ago
|
Add PUSH_CONSTANT_0,1,2; PUSH_0,1; PUSH_NIL bytecodes |
|
More...
|
almost 4 years ago
|
Merge PR #78: Exclude default branch in interpreter loop from compilation |
|
More...
|
almost 4 years ago
|