aboutsummaryrefslogtreecommitdiff
path: root/lib/Bytecode
AgeCommit message (Expand)Author
2007-04-07Avoid a useless temporary constrution.Reid Spencer
2007-04-07Fix another PathWithStatus issue.Reid Spencer
2007-04-04For PR1302:Reid Spencer
2007-03-29For PR789:Reid Spencer
2007-03-29the bytecode reader supports dematerializeFunctionChris Lattner
2007-03-29For PR789:Reid Spencer
2007-03-01Use a simpler constructor for ConstantInt.Reid Spencer
2007-02-28Implement reading of arbitrary precision integers.Reid Spencer
2007-02-28Implement writing of arbitrary precision integers.Reid Spencer
2007-02-15For PR1195:Reid Spencer
2007-02-15For PR1195:Reid Spencer
2007-02-14From Dan Gohman:Chris Lattner
2007-02-13Use a SmallVector to reduce heap traffic. This speeds up bcreader 10%Chris Lattner
2007-02-13now that we can pass ranges into CallInst ctors, eliminate vector heap trafficChris Lattner
2007-02-13stop passing vector into ctorsChris Lattner
2007-02-12avoid creating a temporary string when reading the symbol table for aChris Lattner
2007-02-12Switch ValueSymbolTable to use StringMap<Value*> instead of std::map<std::str...Chris Lattner
2007-02-11Undo the last patch. It slowed down bcwriter by 1.8% but only sped upReid Spencer
2007-02-10Compaction tables don't exist any more.Reid Spencer
2007-02-10For PR1194:Reid Spencer
2007-02-10Fix a comment.Reid Spencer
2007-02-10Change the table datastructure to be a vector<smallvector>, instead ofChris Lattner
2007-02-10Make BytecodeWriter::outputValueSymbolTable *significantly* less abusiveChris Lattner
2007-02-10make the datastructure used in BytecodeWriter::outputValueSymbolTableChris Lattner
2007-02-10Switch typemap over to DenseMap. No significant speedup.Chris Lattner
2007-02-10Switch NodeMap from an std::map to a DenseMap. This speeds up bcwritingChris Lattner
2007-02-10use typedefs where appropriateChris Lattner
2007-02-10The ModuleLevel vector is often quite sparse. Switch it to a DenseMap. ThisChris Lattner
2007-02-10Make the ModuleLevel datastructure more sane. When a function-local valueChris Lattner
2007-02-10Only compute the module levels info once per module, instead of onceChris Lattner
2007-02-10Clone and specialize CreateSlotIfNeeded into CreateFunctionValueSlot to handleChris Lattner
2007-02-10make getSlot/getTypeSlot inlineChris Lattner
2007-02-10getTypeSlot can never failChris Lattner
2007-02-10getSlot can never fail. Make it assert internally, eliminate checks inChris Lattner
2007-02-10simplify getOrCreateTypeSlot, eliminat doInsertType. Eliminate post-order it...Chris Lattner
2007-02-10simplify and speed up recursive type processing.Chris Lattner
2007-02-10rename getOrCreateSlot -> CreateSlotIfNeeded. Noone cares about the retvalChris Lattner
2007-02-10refactor callers of insertType. inline insertType into its one remaining cal...Chris Lattner
2007-02-10simplify code.Chris Lattner
2007-02-10simplify some logic, reduce nestingChris Lattner
2007-02-10Remove dead ctorChris Lattner
2007-02-10ModuleContainsAllFunctionConstants is always trueChris Lattner
2007-02-10only one client of getOrCreateSlot can pass a void typed value. Check typeChris Lattner
2007-02-10inline hasNullValue, rename some variables, simplify some code.Chris Lattner
2007-02-10merge insertValue into its single caller, eliminate some redundant checks.Chris Lattner
2007-02-10merge doInsertValue into insertValueChris Lattner
2007-02-10insertvalue's second operand is always falseChris Lattner
2007-02-10remove dead 'dontIgnore' flag for insertTypeChris Lattner
2007-02-10rename getSlot -> getSlotType and getOrCreateSlot ->getOrCreateTypeSlotChris Lattner
2007-02-09Use int32_t and uint32_t to hopefully help weak compilers (cygwin) notReid Spencer
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2026-02-11 09:06:54 +0000