[PATCH] D61772: [WebAssembly] Don't assume that strongly defined symbol are DSO-local

Fri May 10 10:21:08 PDT 2019

sbc100 added a comment.

In D61772#1498187 <https://reviews.llvm.org/D61772#1498187>, @kripken wrote:

> This may have broken CI, as it shows as one of the commits in the first broken roll here:
>
> https://chromium-review.googlesource.com/c/emscripten-releases/+/1604139
>
> other.test_binaryen_metadce_hello is broken there. Maybe we just need to update expectations? Looks like they rose though, which is odd?
>
> Waterfall CI also shows wasm2.test_dylink_raii_exceptions is broken on Module.dynCall_ifdi is not a function which is maybe also related to this, but just a guess.

There are a couple of different issues right now with the waterfall.

The test_binaryen_metadce_hello failure is very strange.  I've put this fix for it, but I can't see how it broke, certainly doesn't seem to be this change that did it: https://github.com/emscripten-core/emscripten/pull/8581

The test_dylink_raii_exceptions failure has a solution in binaryen: https://github.com/WebAssembly/binaryen/pull/2095.  However I'm not sure that test is still failing on ToT.

Repository:
  rLLD LLVM Linker

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D61772/new/

https://reviews.llvm.org/D61772