clangdev 17.0.1 #241
clangdev 17.0.1 #241
Conversation
This reverts commit 0b6e049.
…nda-forge-pinning 2023.07.18.22.54.11
17.0.0.dev0
…nda-forge-pinning 2023.07.28.20.44.13
automerged PR by conda-forge/automerge-action
…nda-forge-pinning 2023.08.09.23.48.41
clangdev 17.0.0.rc2
automerged PR by conda-forge/automerge-action
h-vetinari
requested review from
JohanMabille,
SylvainCorlay,
chrisburr,
inducer,
isuruf,
jakirkham,
katietz,
timsnyder and
xhochy
as code owners
|
Hi! This is the friendly automated conda-forge-linting service. I just wanted to let you know that I linted all conda-recipes in your PR ( |
This reverts commit a1ba32e.
…nda-forge-pinning 2023.09.19.07.30.00
|
@conda-forge/clangdev, this is ready for review, PTAL. There aren't many relevant changes, biggest issue is dealing with old Macos now that new libcxx requires >10.13 (this will become evident once we have libcxx 17 in the same channel; now main) |
recipe/conda_build_config.yaml
Outdated
| # clang 17 should remain compatible with libcxx 17 (which will be pulled in | ||
| # as a test dependency despite requiring `__osx>=10.13`). Since libcxx has | ||
| # removed workarounds for MacOS<10.13, clang will fail while including | ||
| # the C++ headers, because those rely on things not present in older SDKs. | ||
| MACOSX_SDK_VERSION: # [osx and x86_64] | ||
| - "10.13" # [osx and x86_64] | ||
| MACOSX_DEPLOYMENT_TARGET: # [osx and x86_64] | ||
| - "10.13" # [osx and x86_64] | ||
|
|
There was a problem hiding this comment.
I don't understand why this is needed here. Can you explain?
There was a problem hiding this comment.
The comment tries to lay out my understanding. It was necessary during the rc-phase once there was a libcxx 17, which would get pulled in by clang, which would then fail when including the C++ stdlib (because how libcxx has removed fallbacks for the old macos SDK).
In theory, the requirement __osx>=10.13 on libcxx should be enough to exclude it in this feedstock, but that's not how our CI seems to work. So I needed to bump to 10.13 here as well. I didn't distinguish between SDK and deployment target, because I thought it would be easier on the resolver if the compiler itself also has the requirement, rather than just the stdlib (which is why I had also added it to the compiler activation).
I'm fine to revert this for now if you prefer. We'll just have to re-fix this feedstock one way or another as soon as libcxx 17 is live.
There was a problem hiding this comment.
Let's remove it for now. Let's deal with it when I see an actual error.
This reverts commit 056e5b5.
…nda-forge-pinning 2023.09.26.15.03.47
2 participants
Just a simple bump from what's already in the
rc-channel.Waiting for builds of conda-forge/llvmdev-feedstock#238