Alon Zakai
Alon Zakai
I think if have the CI resources to also test 11 stable, it makes sense to experiment with this. But I think it should be an experiment we are ok...
I'm worried about picking a subset, if we want to claim we support those binaries officially (do we?). Maybe we can check them less frequently or something, though?
Last two comments sgtm (not claim full support, use the FYI bot).
To expand a little on something mentioned earlier without detail: As an approximation to this, I think we can find the LLVM commit closest to a particular release, and then...
Yeah, a one-off test (here on github CI). Then we can officially "bless" that LLVM version and recommend it in our docs, "if you have LLVM release 9, use it...
What's the plan for when a diverging patch lands on LLVM trunk? For example we may get it to emit invokes differently, or you may change the dynamic linking ABI....
> But divergences should be fairly rare and I imagine will be often be linked to features that we can simply not support with the stable version. e.g we could...
We can tag now, but it won't match, I think? Emscripten is already on newer LLVM code after the branch. We'd need to go back in history to find where...
Thanks @tlively ! Then the next step here could be to create a PR here that downloads 1.39.20, then replaces that LLVM with a downloaded LLVM 11 release. If that...
@tlively Ah, I see 😄 Looks like that checks for an emscripten-style table import, and if it exists, reuses the table. And we do have that in our test suite......