Make sure that macros that didn't pass LHS checking are not expanded.

[LeoTestard]

This avoid duplicate errors for things like invalid fragment specifiers, or
parsing errors for ambiguous macros.

[rust-highfive]

r? @nrc

(rust_highfive has picked a reviewer for you, use r? to override)

[oli-obk]

Can you add a compile-fail test for this?

[sanxiyn]

See #29231.

[LeoTestard]

Added two tests, also made sure that the help message indicating which fragment specifiers are valid is still emitted. This required to make OnFail able to span help messages.

[nrc]

r? @pnkfelix

[oli-obk]

lhs should be tt?

[LeoTestard]

Right, looks like I copy-pasted too fast...

[pnkfelix]

I don't know if we have code formatting conventions for macro_rules but I think it looked better before (more uniform in terms of identifying the left and right hand sides of each rule )

[LeoTestard]

:(

[pnkfelix]

(To be clear I am not saying you must change if; I don't even care enough to look up whether we have established convention here)

[pnkfelix]

r=me once above comments are addressed

[LeoTestard]

@pnkfelix Fixed the test to include the help message.

[pnkfelix]

@LeoTestard Travis says you've put the expected error on the wrong line in your test (that, or you need to adjust the span you use in the error reporting)

[LeoTestard]

Actually the span is fine, the error was in the test.

[LeoTestard]

r? @pnkfelix

[pnkfelix]

@bors r+

[bors]

📌 Commit 7d52144 has been approved by pnkfelix