Pass CARGO_HOME on more invocations of cargo

[glasserc]

Fixes #61192.

Motivation for this change

Although this failure only occurs when using Cargo nightly, which isn't actually packaged here yet, this will eventually become a problem when nightly becomes stable.

I'm not really sure that this is the right place for this change. Is the existing absence of CARGO_HOME meant to flush out reproducibility failures when cargo tries to download a package? Or was it just because previously cargo didn't need to do anything in the build or test phases so it wasn't necessary?

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nix-review --run "nix-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Assured whether relevant documentation is up to date
• Fits CONTRIBUTING.md.

---

[samlich]

I lost the log, but I had an issue compiling cargo with it set to $(pwd) as it used two different directories at different times, using $TMPDIR worked.
It also needs to be set for the rustc compilation.

--- a/pkgs/development/compilers/rust/rustc.nix
+++ b/pkgs/development/compilers/rust/rustc.nix
@@ -54,6 +54,8 @@ in stdenv.mkDerivation rec {
   # Increase codegen units to introduce parallelism within the compiler.
   RUSTFLAGS = "-Ccodegen-units=10";

+  # Cargo creates locks in `$CARGO_HOME/.cargo/.package-cache`, which must be writeable
+  CARGO_HOME = "$TMPDIR";
+
   # We need rust to build rust. If we don't provide it, configure will try to download it.

[FRidh]

Please cc the Nixpkgs maintainers of cargo/rust in order to get this further.

[glasserc]

I'm not sure who that is, but based on commit history, maybe it's @Mic92 or @Ericson2314 ?

[samlich]

btw this may be fixed by rust-lang/cargo#7149
If it doesn't, I think they do want to fix it upstream.

[Mic92]

I would rather set it to $TMPDIR though. And it only needs to be set once - maybe in the configurePhase?

[samlich]

With nightly 2019-08-06 built with beta 2019-08-06, I no longer have this issue.

[Mic92]

This should be covered by the following commit: 3372343

[glasserc]

I think so too!