4622 Commits

Author SHA1 Message Date
bors
4c78479610 Auto merge of - RalfJung:miri-as-rustc, r=RalfJung
cargo-miri: never invoke rustc

Always go through 'MIRI_BE_RUSTC=1 miri' instead. This is based on @oli-obk's great idea to add a way to make Miri behave like rustc, which already helped us in https://github.com/rust-lang/miri/pull/1405. Now it means in `cargo-miri` we run *all* crates through the same binary, and use the env var to determine if we compile or interpret them. This makes sure the compiler is consistent.

The `rustc` binary of the current toolchain is now not used at all, only the `miri` binary is. In particular this means we can kill the sysroot consistency check. :)
2020-05-10 17:00:14 +00:00
Ralf Jung
845b89c236 we do not need to set RUSTFLAGS for xargo any more as miri-as-rustc already uses these flags 2020-05-10 18:56:38 +02:00
Ralf Jung
12114c5137 fix dead link in README 2020-05-10 18:52:21 +02:00
Ralf Jung
81046fa5e5 cargo-miri: never invoke rustc, always go through 'MIRI_BE_RUSTC=1 miri' instead 2020-05-10 18:41:03 +02:00
bors
4cbe1f4184 Auto merge of - rust-lang:alloc_map_unlock, r=RalfJung
Update to rustc changes

changes needed once https://github.com/rust-lang/rust/pull/71508 gets merged
2020-05-10 08:32:58 +00:00
Ralf Jung
4f06197aff rustup 2020-05-10 10:32:29 +02:00
Oliver Scherer
3fdab9c446 Update to rustc changes 2020-05-10 10:14:03 +02:00
bors
8a6396ffa5 Auto merge of - RalfJung:stage-0, r=RalfJung
make Miri work in rustc bootstrap stage 0

Fixes https://github.com/rust-lang/rust/issues/52856
2020-05-09 12:09:15 +00:00
Ralf Jung
e65d87b110
Typo 2020-05-09 14:08:58 +02:00
Ralf Jung
ac65350789 adjust default sysroot when being rustc
Also while at it, refactor how we pass the default Miri flags
2020-05-09 13:42:07 +02:00
Ralf Jung
ba801a45db make Miri work in rustc bootstrap stage 0 2020-05-09 13:19:23 +02:00
bors
4eaf05c4ec Auto merge of - RalfJung:cargo-miri-host-detect, r=RalfJung
re-do cargo-miri host/target detection logic to match rustbuild

@oli-obk I think that's better than looking at `--emit` like we did before... what do you think?
2020-05-09 10:06:04 +00:00
Ralf Jung
024cc435f4 avoid env::var which requires valid UTF-8 2020-05-09 12:04:21 +02:00
Ralf Jung
20097be2fe more comment 2020-05-09 11:54:45 +02:00
Ralf Jung
1ba42b9f55
Wording
Co-authored-by: Oliver Scherer <github35764891676564198441@oli-obk.de>
2020-05-09 11:53:24 +02:00
Ralf Jung
e73fc97f0b cargo-miri: honor RUSTC env var 2020-05-09 11:52:26 +02:00
Ralf Jung
70f83a342c re-do cargo-miri host/target detection logic to match rustbuild 2020-05-09 11:45:43 +02:00
bors
17f740e9d3 Auto merge of - RalfJung:rustup, r=RalfJung
bump Rust, fix for renames
2020-05-09 08:15:45 +00:00
Ralf Jung
379ac82a1c bump Rust, fix for renames 2020-05-09 10:15:09 +02:00
bors
95dec9cee4 Auto merge of - RalfJung:cargo-miri-help, r=RalfJung
fix cargo-miri intercepting --help/--version arguments

With this, `cargo miri test -- -- --help` should behave as expected and pass `--help` to the test runner.

Thanks to @brson for the report.
2020-05-08 07:56:38 +00:00
Ralf Jung
914e483c89 fix cargo-miri intercepting --help/--version arguments 2020-05-08 09:55:28 +02:00
bors
7f3366288d Auto merge of - RalfJung:rustup, r=RalfJung
rustup
2020-05-07 21:59:57 +00:00
Ralf Jung
b4ad90669b rustup 2020-05-07 23:59:26 +02:00
bors
c49e97d9e7 Auto merge of - brson:tikv-2, r=RalfJung
Add two more TiKV bugs to trophy case
2020-05-07 09:22:17 +00:00
Ralf Jung
132feb88db
adjust wording 2020-05-07 11:21:31 +02:00
bors
2d3261de75 Auto merge of - RalfJung:rustup, r=RalfJung
rustup; fix error messages
2020-05-07 06:42:59 +00:00
Ralf Jung
fb3a067ac6 rustup; fix error messages 2020-05-07 08:38:59 +02:00
Brian Anderson
356848c9b9 Add two more TiKV bugs to trophy case 2020-05-07 02:42:12 +00:00
bors
308458a4ab Auto merge of - RalfJung:float-cast-tests, r=RalfJung
copy some float cast tests from rustc

The more the merrirer!
2020-05-06 20:47:34 +00:00
Ralf Jung
da6846c8a9 copy some float cast tests from rustc 2020-05-06 22:46:42 +02:00
bors
d2e283bf6a Auto merge of - RalfJung:fs-mode, r=RalfJung
fs shim: check that mode is default
2020-05-05 14:49:48 +00:00
Ralf Jung
45ef97535f fs shim: check that mode is default 2020-05-05 16:49:01 +02:00
bors
13f1b5f653 Auto merge of - RalfJung:arg-count-cleanup, r=RalfJung
Make sure we check argument count everywhere and argument size for non-stub shims
2020-05-05 10:21:46 +00:00
Ralf Jung
40800cfa19 make sure we check the size of all arguments 2020-05-05 12:19:54 +02:00
Ralf Jung
5656cb73d4 fix a comment now that we have concurrency 2020-05-05 11:59:22 +02:00
Ralf Jung
cd6be98852 make sure we check argument count everywhere 2020-05-05 11:59:22 +02:00
Ralf Jung
5566e3901c deduplicate FD extraction 2020-05-05 11:44:33 +02:00
bors
2f974f6e1f Auto merge of - toc-the-younger:check-arg-count, r=RalfJung
Check that shims are called with the correct number of arguments

Fixes 

Some shims don't use all their arguments, I've assumed they are being called in tests with the correct number of arguments here.
2020-05-05 09:08:10 +00:00
Chase Albert
f741f2cc71 Correct the test. 2020-05-04 23:22:00 -04:00
Chase Albert
4e3d1fee51 Address comments. 2020-05-04 14:24:22 -04:00
Chase Albert
f46d197431 Test for too many args. 2020-05-04 13:51:23 -04:00
Chase Albert
4d3dff2add Address consistency of naming for unused/merely validated arguments. 2020-05-04 13:51:23 -04:00
Chase Albert
46aaab30fe Add a test for check_arg_count. 2020-05-04 13:51:23 -04:00
Chase Albert
78f329513a Check that shims are called with the correct number of arguments 2020-05-04 13:51:21 -04:00
bors
a49234ebaf Auto merge of - RalfJung:win-lock-check, r=RalfJung
Windows lock primitives: check that we are truly single-threaded

Also rustup because why not.
2020-05-04 07:45:35 +00:00
Ralf Jung
ec95ed4556 rename single-threaded sync test 2020-05-04 09:45:15 +02:00
Ralf Jung
65c3914a29 rustup 2020-05-04 09:43:09 +02:00
Ralf Jung
61a86e1ffe Windows lock primitives: check that we are truly sinle-threaded 2020-05-04 09:37:40 +02:00
bors
1e87c9f8ba Auto merge of - RalfJung:conc, r=RalfJung
Make spin_loop yield, and a panicking thread test

@vakaras agreed it would make sense to make `spin_loop` yield, so this implements that.

Also add a test for a panic being raised in one thread, caught by the panic machinery, and then expose through the join handle. Miraculously, that just works. :)
2020-05-03 11:23:47 +00:00
Ralf Jung
6b18f6e106 fix setting thread name on macOS 2020-05-03 13:22:56 +02:00