[cling] Enable cling to be built with Emscripten

[mcbarton]

This Pull request:

Changes or fixes:

This change enables cling to be built with Emscripten. The Emscripten build of cling doesn't currently work, but these changes allow it to build. If this goes in I will begin debugging what changes are needed to enable it to function in a web browser and node through CppInterOp, which has an Emscripten ci (including tests).

Checklist:

• tested changes locally
• updated the docs (if necessary)

This PR fixes #

[ferdymercury]

Maybe related: #20225

[github-actions]

Test Results

    22 files      22 suites   3d 14h 27m 50s ⏱️
 3 705 tests  3 705 ✅ 0 💤 0 ❌
79 559 runs  79 559 ✅ 0 💤 0 ❌

Results for commit 62f6ac7.

[devajithvs]

I think the changes in #20225 should fix Cling build with Emscripten. Let's wait for that before making this change.

[hahnjo]

In my opinion, #20225 is the more correct approach because it's not about building with Emscripten, but really about the missing NVPTX backend.

[vgvassilev]

In my opinion, #20225 is the more correct approach because it's not about building with Emscripten, but really about the missing NVPTX backend.

What has to do the emscripten work with some other PR we discussed. Is there a reason to block the current PR?

[hahnjo]

In my opinion, #20225 is the more correct approach because it's not about building with Emscripten, but really about the missing NVPTX backend.

What has to do the emscripten work with some other PR we discussed. Is there a reason to block the current PR?

I don't understand the first sentence of the comment. And yes, I object to this PR because #ifndef __EMSCRIPTEN__ is wrong and #if LLVM_HAS_NVPTX_TARGET from #20225 is correct.

[vgvassilev]

and #if LLVM_HAS_NVPTX_TARGET from #20225 is correct.

In your opinion.

Now I see what you meant. Yes, we should not wrap the nvptx initialization around emscripten clauses.

[hahnjo]

and #if LLVM_HAS_NVPTX_TARGET from #20225 is correct.

In your opinion.

Yes, obviously. You asked for the reason to block this PR, there it is. I'm open to discussion, but that's the technical arguments I see so far.

[vgvassilev]

and #if LLVM_HAS_NVPTX_TARGET from #20225 is correct.

In your opinion.

Yes, obviously. You asked for the reason to block this PR, there it is. I'm open to discussion, but that's the technical arguments I see so far.

Apologies -- I overlooked. Updated my comment...

[mcbarton]

Closing as I checked locally and #20225 achieves the same result of allowing cling to be compiled with Emscripten.