Update xeus-cpp to CppInterOp 1.8.0#436
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #436 +/- ##
==========================================
- Coverage 81.96% 81.25% -0.72%
==========================================
Files 21 21
Lines 959 864 -95
Branches 87 78 -9
==========================================
- Hits 786 702 -84
+ Misses 173 162 -11
... and 14 files with indirect coverage changes
🚀 New features to boost your workflow:
|
anutosh491
left a comment
There was a problem hiding this comment.
Please wait for sometime. I need to find the time to verify the cppinterop build on emscripten-forge !
06c0404 to
f4b789a
Compare
|
See the build was wrong, using llvm 20.1.8 instead of 21.1.8 (https://github.com/emscripten-forge/recipes/actions/runs/21440156682/job/61740836429#step:7:117) Let's fix this ! |
conda-forge/cppinterop-feedstock#27 ? But is that blocking this PR? |
This is because the Emscripten forge bot which updated the CppInterOp version didn't update the llvm version (see https://github.com/emscripten-forge/recipes/pull/4332/changes ). We need the llvm version updated like the PR I had opened here https://github.com/emscripten-forge/recipes/pull/4332/changes (but opened in a fresh PR) , but this must come after your resource headers PR is in https://github.com/emscripten-forge/recipes/pull/4332/changes. |
|
@vgvassilev conda-forge/cppinterop-feedstock#27 updates everything in the cppinterop feedstock to llvm 21 (things which should have gone in the PR which updated to version 1.8.0) . I suspect none of the changes in it effect the ci passing, but its best it goes in. |
5 mins more and we should have a fixed cppinterop build on emscripten-forge (the build number should be |
fed9b61 to
17d0a62
Compare
anutosh491
left a comment
There was a problem hiding this comment.
I tried understanding the unpinning on the cxx-compiler. As far as I remember that wasn't something that just work right off the box. But if it does #186 can now be closed !
Curious if this started working too compiler-research/CppInterOp#422
I'll get back to this later if it deserves any attention ! Rest everything looks OK
|
I guess we wait for the cppinterop build after the feedstock merge to be ready -> then rerun the jobs here and then merge ?! |
|
The new conda-forge package is being used by the ci (see https://github.com/compiler-research/xeus-cpp/pull/436/checks#step:3:103 for an example showing this). The ci has passed. Merging. |
|
I was waiting for your fix on the feedstock "the llvm 21 fix" to land and then we could have had a build against that in the CI and then merged. But ig its fine |
The ci did build against that feedstock package you mention. That is what I was pointing to when I gave the link https://github.com/compiler-research/xeus-cpp/pull/436/checks#step:3:103 |
| - nlohmann_json | ||
| - nlohmann_json-abi | ||
| - CppInterOp | ||
| - CppInterOp>1.7.0 |
There was a problem hiding this comment.
In don't understand the strict upper bound on 1.7.0.
Shouldn't it be a non-strict on 1.8?
This PR gets xeus-cpp ready for the impending release of CppInterOp.
Fixes #186