Fix data-layout field in x86_64-unknown-linux-gnu.json test file

[phil-opp]

The current data-layout causes the following error:

rustc: /checkout/src/llvm/lib/CodeGen/MachineFunction.cpp:151: void llvm::MachineFunction::init(): Assertion `Target.isCompatibleDataLayout(getDataLayout()) && "Can't create a MachineFunction using a Module with a " "Target-incompatible DataLayout attached\n"' failed.

The new value was generated according to this comment by @japaric.

[carols10cents]

r? @japaric

Thanks for the PR! We’ll periodically check in on it to make sure that @japaric or someone else from the team reviews it soon.

[japaric]

The change LGTM but the data-layout changed quite long time ago; why hasn't this break a test?

It seems to me that this file is not being used at all. When rustc sees --target $T it will use the built-in definition of $T if there is one; otherwise it will look for a specification file named $T.json. So I think that the test that is meant to use this file is actually using the built-in target definition. I think it would be better to rename the file to not match the name of a built-in target and then adjust the test accordingly.

Could you look into it @phil-opp? (TBH, I don't even know which test is supposed to be using this file.)

[phil-opp]

the data-layout changed quite long time ago; why hasn't this break a test?

Yeah, that wondered me too. I think that the file should be used by the target-specs test (see the Makefile).

I added a commit that renames the file to my-x86_64-unknown-linux-gnu-platform.json (to match the other test targets) and adjusts the target in the Makefile.

(I couldn't run the test locally, since I only have my laptop with a small SSD at the moment.)

[japaric]

Thanks, @phil-opp.

Travis seems to be happy so let's land this.

@bors r+ rollup

[bors]

📌 Commit 06b9168 has been approved by japaric