sccache v0.2.13 does not work when built with rustc 1.48

[SomeoneToIgnore]

The same dummy "hello world" project + sccache does not work for rustc 1.48 but works for 1.47.

Here's a minimal example:

Create the script locally:

#! /usr/bin/env bash

set -e
set -x

env USER=test-user cargo new test-project --bin
cd test-project
echo "Running without sccache"
cargo run
cargo clean
cargo install sccache
echo "Running with sccache"
RUSTC_WRAPPER="/usr/local/cargo/bin/sccache" cargo run
sccache -s

and then run it in Docker:

docker run --rm -v $(pwd)/test-sccache.sh:/test-sccache.sh rust:1.47 /test-sccache.sh

docker run --rm -v $(pwd)/test-sccache.sh:/test-sccache.sh rust:1.48 /test-sccache.sh

Expected: both runs work identically and finish successfully

Actual: 1.47 version works as expected, 1.48 version fails on the RUSTC_WRAPPER="/usr/local/cargo/bin/sccache" cargo run step:

Running with sccache
+ echo 'Running with sccache'
+ RUSTC_WRAPPER=/usr/local/cargo/bin/sccache
+ cargo run
error: failed to run `rustc` to learn about target-specific information

Caused by:
  process didn't exit successfully: `/usr/local/cargo/bin/sccache rustc - --crate-name ___ --print=file-names --crate-type bin --crate-type rlib --crate-type dylib --crate-type cdylib --crate-type staticlib --crate-type proc-macro --print=sysroot --print=cfg` (exit code: 2)
  --- stderr
  error: failed to execute compile
  caused by: Failed to send data to or receive data from server
  caused by: Failed to read response header
  caused by: failed to fill whole buffer

[drahnr]

@SomeoneToIgnore going to look into that later tonight/early tomorrow.

Could you check if the latest master works though? And see which unit tests fail?

Looking into the release notes, there are some potential cli flag changes that could cause the failure https://github.com/rust-lang/rust/releases/tag/1.48.0

[SomeoneToIgnore]

It does work actually, thanks for the tip.

So it's the release that we need.

[drahnr]

Unfortunately, yes - but I'd recommend to read #876

[bossmc]

In case people are interested this looks like a breaking (actually the old behaviour was UB so this isn't breaking) change in rust 1.48 (rust-lang/rust#71274) where mem::uninitialized now panics if the type can't hold the zero value:

$ SCCACHE_LOG=debug SCCACHE_START_SERVER=1 SCCACHE_NO_DAEMON=1 target/debug/sccache
thread 'main' panicked at 'attempted to leave type `linked_hash_map::Node<std::ffi::OsString, u64>` uninitialized, which is invalid', /home/andy/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/lib/rustlib/src/rust/library/core/src/mem/mod.rs:658:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

(I assume OsString is a NonNull<u8> or similar).

Until (unless?) Mozilla cut a new release there is a pre-built binary alongside each release (e.g. at https://github.com/mozilla/sccache/releases/tag/0.2.13) that works, or you can use master or @drahnr's fork at https://github.com/paritytech/sccache.

[glandium]

Duplicate of #875