Rollup merge of #65321 - Mark-Simulacrum:remove-linker-no-utf8-test, r=nikomatsakis
Remove painful test that is not pulling its weight Research suggests that we are not properly testing this case anyway, and even if we were, it is unlikely that we will regress here -- or, perhaps more accurately, if we do, I am uncertain that we care too much. It definitely seems like an edge case, and one that is particularly unlikely to occur as time goes on. Fixes https://github.com/rust-lang/rust/issues/63520
This commit is contained in:
commit
b93203ff07
3 changed files with 0 additions and 39 deletions
|
@ -1,23 +0,0 @@
|
|||
-include ../tools.mk
|
||||
|
||||
# Make sure we don't ICE if the linker prints a non-UTF-8 error message.
|
||||
|
||||
# ignore-windows
|
||||
#
|
||||
# This does not work in its current form on windows, possibly due to
|
||||
# gcc bugs or something about valid Windows paths. See issue #29151
|
||||
# for more information.
|
||||
|
||||
# ignore-macos
|
||||
#
|
||||
# This also does not work on Apple APFS due to the filesystem requiring
|
||||
# valid UTF-8 paths.
|
||||
|
||||
# The zzz it to allow humans to tab complete or glob this thing.
|
||||
bad_dir := $(TMPDIR)/zzz$$'\xff'
|
||||
|
||||
all:
|
||||
$(RUSTC) library.rs
|
||||
mkdir $(bad_dir)
|
||||
mv $(TMPDIR)/liblibrary.a $(bad_dir)
|
||||
$(RUSTC) -L $(bad_dir) exec.rs 2>&1 | $(CGREP) this_symbol_not_defined
|
|
@ -1,6 +0,0 @@
|
|||
#[link(name="library")]
|
||||
extern "C" {
|
||||
fn foo();
|
||||
}
|
||||
|
||||
fn main() { unsafe { foo(); } }
|
|
@ -1,10 +0,0 @@
|
|||
#![crate_type = "staticlib"]
|
||||
|
||||
extern "C" {
|
||||
fn this_symbol_not_defined();
|
||||
}
|
||||
|
||||
#[no_mangle]
|
||||
pub extern "C" fn foo() {
|
||||
unsafe { this_symbol_not_defined(); }
|
||||
}
|
Loading…
Add table
Reference in a new issue