Add support for llvm.sin.f(32|64) intrinsics

[Maher4Ever]

This adds support for the llvm.sin.f(32|64) intrinsic to Emscripten. More details can be found in the issue #4338. The accompanying change to fastcomp is in pull request emscripten-core/emscripten-fastcomp#143

[kripken]

Looks good, thanks.

Please rebase (and then push -f to this branch) to get rid of the merge commit.

Also please add a test. Can grep for the similar functions next to it, in tests/, to find the relevant test that I think you can just add a few lines in, if I remember correctly.

[Maher4Ever]

Thanks about the rebase tip. It makes the history so much cleaner!

I greped for each of the following strings inside tests/ but found nothing:

• llvm_log and llvm.log
• llvm_sqrt and llvm.sqrt
• llvm_pow and llvm.pow
• llvm_exp and llvm,exp

Are you aware of a more specific folder or file where I should find the tests?

[kripken]

Looks like the relevant tests are in tests/core/test_llvm_intrinsics.in and .out.

[Maher4Ever]

Done :)

[kripken]

Thanks!