Attempt to fix tests on master #662
Conversation
alexcrichton
force-pushed
the
fix-tests
branch
from
c9c1878 to
eeaef70
Compare
* Make all doctests use items from the real `std` rather than this crate, it's just easier * Handle debuginfo weirdness by flagging functions as `no_mangle` that we're looking for instructions within.
alexcrichton
force-pushed
the
fix-tests
branch
from
eeaef70 to
5353052
Compare
alexcrichton
force-pushed
the
fix-tests
branch
from
9c365bd to
e463354
Compare
|
FWIW I believe a number of these changes are motivated by the recent LLVM upgrade in rust-lang/rust, others are just minor things to accrue over time I believe |
| # extern crate core_arch as std; | ||
| # #[cfg(not(dox))] | ||
| # #[macro_use(is_x86_feature_detected)] | ||
| # extern crate std_detect; |
There was a problem hiding this comment.
Should we be doing this everywhere ? There are many doc examples where we do this dance. EDIT: I see some more of this below, sorry about the noise!
EDIT2: so grepping for "real_std" reveals still some occurrences of this. I am not sure exactly why this change was needed here but not everywhere else.
Opened: #665
There was a problem hiding this comment.
Yeah I don't really know what was happening to have the doctests fail, but this is all in std now so none of this was necessary any more
| @@ -99,6 +99,7 @@ pub fn assert_instr( | |||
| let shim_name_str = format!("{}{}", shim_name, assert_name); | |||
| let to_test = quote! { | |||
| #attrs | |||
| #[no_mangle] | |||
There was a problem hiding this comment.
We probably should add a comment around here about which exact problem #[no_mangle] solves (debug-info weirdness).
EDIT: opened #664
stdrather than thiscrate, it's just easier
no_manglethatwe're looking for instructions within.