The Code Extractor does not provide an easy mechanism for determining the inputs and outputs after extraction has occurred, this patch gives the ability to pass in empty SetVectors to be filled with the inputs and outputs if they need to be analyzed.
Details
Diff Detail
Unit Tests
Time | Test | |
---|---|---|
14,070 ms | x64 debian > libarcher.races::task-dependency.c |
Event Timeline
LGTM with documentation nit.
llvm/include/llvm/Transforms/Utils/CodeExtractor.h | ||
---|---|---|
142 | Can you add documentation for Inputs and Outputs explaining that they are out-params? |
llvm/unittests/Transforms/Utils/CodeExtractorTest.cpp | ||
---|---|---|
118 | Can all of these use EXPECT_EQ? |
LGTM with one more nit
llvm/include/llvm/Transforms/Utils/CodeExtractor.h | ||
---|---|---|
147 | maybe just say "filled with input values to the outlined function"? even though this is a typedef of SetVector, it's a little confusing when looking at this locally. (similar nit with Outputs) |
Can you add documentation for Inputs and Outputs explaining that they are out-params?