Relooper CFG optimizations (#1759)

Previously the relooper would do some optimizations when deciding when to use an if vs a switch, how to group blocks, etc. This PR adds an additional pre-optimization phase with some basic but useful simplify-cfg style passes,

* Skip empty blocks when they have just one exit.
* Merge exiting branches when they are equivalent.
* Canonicalize block contents to make such comparisons more useful.
* Turn a trivial one-target switch into a simple branch.

This can help in noticeable ways when running the rereloop pass, e.g. on LLVM wasm backend output.

Also:

* Binaryen C API changes to the relooper, which now gets a Module for its constructor. It needs it for the optimizations, as it may construct new nodes.
* Many relooper-fuzzer improvements.
* Clean up HashType usage.

1 files changed, 2 insertions, 2 deletions

diff --git a/test/example/relooper-fuzz.c b/test/example/relooper-fuzz.c
index 2b8744d5e..606274683 100644
--- a/test/example/relooper-fuzz.c
+++ b/test/example/relooper-fuzz.c
@@ -62,7 +62,7 @@ int main() {
 
   // contents of main() begin here
 
-  RelooperRef relooper = RelooperCreate();
+  RelooperRef relooper = RelooperCreate(module);
 
 
   RelooperBlockRef b0;
@@ -223,7 +223,7 @@ int main() {
 
   RelooperAddBranch(b8, b4, NULL, NULL);
 
-  BinaryenExpressionRef body = RelooperRenderAndDispose(relooper, b0, 1, module);
+  BinaryenExpressionRef body = RelooperRenderAndDispose(relooper, b0, 1);
 
   int decisions[] = { 89, 12, 78, 149, 118, 179, 127, 80, 21, 34, 119, 98, 38, 29, 36, 147, 13, 55, 166, 16, 143, 52, 130, 150, 176, 91, 34 };
   int numDecisions = sizeof(decisions)/sizeof(int);