While usage of change_dir_locked is synchronized against itself, it's not
synchronized against other relative path usage, so I'm of the opinion that it
just really doesn't help in running tests. In order to prevent the problems that
have been cropping up, this completely removes the function.
All existing tests (except one) using it have been moved to run-pass tests where
they get their own process and don't need to be synchronized with anyone else.
There is one now-ignored rustpkg test because when I moved it to a run-pass test
apparently run-pass isn't set up to have 'extern mod rustc' (it ends up having
linkage failures).
While usage of change_dir_locked is synchronized against itself, it's not
synchronized against other relative path usage, so I'm of the opinion that it
just really doesn't help in running tests. In order to prevent the problems that
have been cropping up, this completely removes the function.
All existing tests (except one) using it have been moved to run-pass tests where
they get their own process and don't need to be synchronized with anyone else.
There is one now-ignored rustpkg test because when I moved it to a run-pass test
apparently run-pass isn't set up to have 'extern mod rustc' (it ends up having
linkage failures).
Servo is hitting this problem, so this is a workaround for lack of a real solution.
No tests because I couldn't actually reproduce the problem with either of the testcases in #9129
Since function pointers do not carry along the function attributes with
them in the type, this needs to be set on the call instruction itself.
Closes#9152
r? @brson As per rustpkg.md, rustpkg now builds in a target-specific
subdirectory of build/, and installs libraries into a target-specific
subdirectory of lib.
Closes#8672
The glob tests cannot change the current working directory because the other tests (namely the fileinput ones) depend on the current working directory not changing.
The normal unit tests cannot change the current working directory because it
messes with the other tests which depend on a particular working directory.
As per rustpkg.md, rustpkg now builds in a target-specific
subdirectory of build/, and installs libraries into a target-specific
subdirectory of lib.
Closes#8672
Since function pointers do not carry along the function attributes with
them in the type, this needs to be set on the call instruction itself.
Closes#9152
These tests are being very flaky on the bots, and the reason is that files are
being created and then when attempted to get read they actually don't exist. I'm
not entirely sure why this is happening, but I also don't fully trust the
std::io implemention using @-boxes to close/flush/write files at the right time.
This moves the tests to using std::rt::io which is hopefully more robust and
something that we can actually reason about. Sadly, due to #8810, these tests
fail on windows, so they're all ignored on windows right now.
Testing this is a little tricky as an intermediate temporary directory is only used for remote git repositories and therefore that path cannot be reliably exercised in the tests.
r? @metajack Install to the first directory in the RUST_PATH if the user set a
RUST_PATH. In the case where RUST_PATH isn't set, the behavior
remains unchanged.
Closes#7402
Install to the first directory in the RUST_PATH if the user set a
RUST_PATH. In the case where RUST_PATH isn't set, the behavior
remains unchanged.
Closes#7402
