0
I Use This!
Inactive

Commits : Listings

Analyzed 1 day ago. based on code collected 1 day ago.
Aug 15, 2024 — Aug 15, 2025
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Make changes to the live range analysis as required after assigning registers, the code cannot be exactly the same as before assigning the registers. More... over 13 years ago
Change the way loop headers and footers were handled in each individual stages, since we already take care of this while constructing the dominator tree for SSA. More... over 13 years ago
Implement the loop detection algorithm using the dominator tree. More... over 13 years ago
Add an explicit function name type for the global symbol table entries so that memory need not be allocated for them. More... over 13 years ago
Fix a styling issue. More... over 13 years ago
Fix the bootstrap file to use linker and elf builder. More... over 13 years ago
Render the binary string for debugging. More... over 13 years ago
Style fixes. More... over 13 years ago
Use the new format of the function offset map and also update the size after each addition of instruction to the binary. More... over 13 years ago
Hard code the values for builtin functions for now, when we have them implemented, we can populate their values. More... over 13 years ago
Add a precompiled section that bootstraps the binary by calling main. More... over 13 years ago
Module that implements pre-compiled functions. More... over 13 years ago
ELF class now gets the linker object from which it can build the entire binary. More... over 13 years ago
What a blunder, adding does a carry, we don't want that More... over 13 years ago
No chr conversion for data, we never wanted that. More... over 13 years ago
Hard code the address values used for the program text and data by defining constants for them for now. More... over 13 years ago
Process all the jump targets within a function and rebuild the jump binaries. More... over 13 years ago
Prologue and Epilogue should get the function names not the instruciton mnemonics. More... over 13 years ago
When we have a main function don't allocate stack for it. More... over 13 years ago
Returns should get a register when there is a value. More... over 13 years ago
Implement the call instruction using the x86_64 call instruction not jump instruction. More... over 13 years ago
Codegenerator should use the new branch instruction format of pointing to branch CFG nodes. More... over 13 years ago
When the result of the phi instruction is not colored, don't deconstruct the phi instruction. More... over 13 years ago
Make optimized removal a set so that we can get set differences and all such set operations when required without compromising on the Big-Oh complexity. More... over 13 years ago
Implement constant folding and dead code elimination by making use of the new branch operands format (that uses nodes). More... over 13 years ago
We don't do this processed edge recording because if there is a two node loop, this code won't render the loop footer to header edge which is painful to debug. More... over 13 years ago
After building the CFG make the branch instructions point to the CFG nodes and not the labels. More... over 13 years ago
Add a separate method to render branch operands to print the node values since branch instructions will be changed to point to the basic block node in the next few commits. More... over 13 years ago
MOV has a different way of handling the registers to immediate operands. More... over 13 years ago
Teach the OPCODE table about the OPCODE Extensions in case of immediate source operands. More... over 13 years ago