-
Notifications
You must be signed in to change notification settings - Fork 12.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
rustc: Update LLVM, remove dead wasm code #58408
Conversation
(rust_highfive has picked a reviewer for you, use r? to override) |
|
I suspect you're going to run into the new toolchain requirements on the dist builders, per: I'm working on that right now for dist-*-linux builders, at least. In the meantime you can set |
☔ The latest upstream changes (presumably #58341) made this pull request unmergeable. Please resolve the merge conflicts. |
59de990
to
3e2d424
Compare
☔ The latest upstream changes (presumably #58406) made this pull request unmergeable. Please resolve the merge conflicts. |
3e2d424
to
fcb7900
Compare
ping @matthewjasper, do you think you'll have a chance to take a look at this? No worries if not, can just r? someone else! |
This isn't an area of the code I know much about. Could you r? someone else please. |
We technically support LLVM 6+ (i.e., that's what the distro LLVM builder in CI uses) -- do we want to continue having the "dead" wasm code? |
@Mark-Simulacrum we don't support the wasm target on anything other than our in-tree LLVM |
@bors r+ Thanks, @alexcrichton! |
📌 Commit fcb79008ce14dd8ab729b1e98f0f306b47c0192c has been approved by |
Since it has submodule updates, ill use this an inbtw rollup filler, @bors p=1 |
Apparently bors interpreted this wrongly... @bors p=1 rollup- |
⌛ Testing commit fcb79008ce14dd8ab729b1e98f0f306b47c0192c with merge a707a104c3fe2fe04958f73922c5f28927ae720d... |
💔 Test failed - status-appveyor |
|
☔ The latest upstream changes (presumably #58728) made this pull request unmergeable. Please resolve the merge conflicts. |
The job Click to expand the log.
I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
Docker rebuilt for that container, taking 2.5h, causing a timeout @bors: retry |
Why is In #58416 this seemed to be a problem too. There it really did need docker updates, of course, but it seemed to take a few tries before they all finally stuck. |
rustc: Update LLVM, remove dead wasm code This commit updates the LLVM branch to the rebased version of the upstream release/8.x branch. This includes a wasm patch which means that the `rewrite_imports` pass in rustc is no longer needed (yay!) and we can instead rely on `wasm-import-module`, an attribute we're already emitting, to take care of all the work.
☀️ Test successful - checks-travis, status-appveyor |
@cuviper it's probably hashing issues like before, although I won't pretend to know what hashing issues it is... In any case I'm surprised this landed! I guess we don't actually have all that many toolchains to update |
I've opened up #58852 to handle the NetBSD toolchain, will look into Windows next. |
Originally added in rust-lang#55935 to test build times, this was reverted in rust-lang#56201 due to a belief that it caused the exit code 259 spurious errors. We've since learned, however, that the 259 exit code is likely not related to this image update as we're getting it in a number of locations now. VS2017 looks like it may be required to compile LLVm in the near future, notably discovered by rust-lang#58408 where we attempted to update LLVM.
And #58854 is the attempt for Windows |
…Simulacrum Update toolchain to build NetBSD release This allows us to remove the "allow old toolchains" flag we pass to LLVM, ensuring that we'll be up to date when LLVM needs us to be! This is a follow-up from rust-lang#58408 where NetBSD was explicitly whitelisted to allow older toolchains.
…-Simulacrum appveyor: Use VS2017 for all our images Originally added in rust-lang#55935 to test build times, this was reverted in rust-lang#56201 due to a belief that it caused the exit code 259 spurious errors. We've since learned, however, that the 259 exit code is likely not related to this image update as we're getting it in a number of locations now. VS2017 looks like it may be required to compile LLVm in the near future, notably discovered by rust-lang#58408 where we attempted to update LLVM.
…-Simulacrum appveyor: Use VS2017 for all our images Originally added in rust-lang#55935 to test build times, this was reverted in rust-lang#56201 due to a belief that it caused the exit code 259 spurious errors. We've since learned, however, that the 259 exit code is likely not related to this image update as we're getting it in a number of locations now. VS2017 looks like it may be required to compile LLVm in the near future, notably discovered by rust-lang#58408 where we attempted to update LLVM.
…-Simulacrum appveyor: Use VS2017 for all our images Originally added in rust-lang#55935 to test build times, this was reverted in rust-lang#56201 due to a belief that it caused the exit code 259 spurious errors. We've since learned, however, that the 259 exit code is likely not related to this image update as we're getting it in a number of locations now. VS2017 looks like it may be required to compile LLVm in the near future, notably discovered by rust-lang#58408 where we attempted to update LLVM.
…-Simulacrum appveyor: Use VS2017 for all our images Originally added in rust-lang#55935 to test build times, this was reverted in rust-lang#56201 due to a belief that it caused the exit code 259 spurious errors. We've since learned, however, that the 259 exit code is likely not related to this image update as we're getting it in a number of locations now. VS2017 looks like it may be required to compile LLVm in the near future, notably discovered by rust-lang#58408 where we attempted to update LLVM.
This commit updates the LLVM branch to the rebased version of the
upstream release/8.x branch. This includes a wasm patch which means that
the
rewrite_imports
pass in rustc is no longer needed (yay!) and wecan instead rely on
wasm-import-module
, an attribute we're alreadyemitting, to take care of all the work.