This overhauls `std::run` to instead run on top of libuv. This is *not* in a mergeable state, I've been attempting to diagnose failures in the compiletest suite. I've managed to find a fair number of bugs so far, but I still don't seem to be done yet.
Notable changes:
* This requires upgrading libuv. From the discussion on #6567, I took libuv master from a few days ago, applied one patch to fix process spawning with multiple event loops in libuv, and pushed to my own fork
* The build system for libuv has changed since we last used it. There's some extra checkout from a google build system which apparently does all the magic if you don't want to require autotools, and the google system just requires python. I updated the Makefile to get this build system and build libuv with it instead. This is untested on windows and arm, and both will probably need to see some improvement.
* This required adding some pipe bindings to libuv as well. Currently the support is pretty simple and probably completely unsafe for pipes, but you at least get read/write methods. This is necessary for capturing output of processes.
* I didn't redesign `std::run` at all, I simply tried to reimplement all the existing functionality on top of libuv. Some functions ended up dying, but nothing major. All uses of `std::run` in the compiler still work just fine.
I'm not quite sure how the rest of the runtime deals with this, but I marked process structures as `no_send` because the waiting/waking up has to happen in the same event loop right now. If processes start migrating between event loops then very bad things can happen. This may be what threadsafe I/O would fix, and I would be more than willing to rebase on that if it lands first.
Anyway, for now I wanted to put this up for review, I'm still investigating the corruption/deadlock bugs, but this is in an *almost* workable state. Once I find the bugs I'll also rebase on the current master.
There were two main differences with the old libuv and the master version:
1. The uv_last_error function is now gone. The error code returned by each
function is the "last error" so now a UvError is just a wrapper around a
c_int.
2. The repo no longer includes a makefile, and the build system has change.
According to the build directions on joyent/libuv, this now downloads a `gyp`
program into the `libuv/build` directory and builds using that. This
shouldn't add any dependences on autotools or anything like that.
Closes#8407Closes#6567Closes#6315
`rustpkg build` et al were only checking one directory up to see if it
was in a dir named "src". Ditch that entirely and instead check if the
cwd is descended from any of the workspace paths. Besides being more
intelligent about whether or not something is a workspace, this also
allows for package ids composed of multiple path components.
r? @catamorphism
`rustpkg build` et al were only checking one directory up to see if it
was in a dir named "src". Ditch that entirely and instead check if the
cwd is descended from any of the workspace paths. Besides being more
intelligent about whether or not something is a workspace, this also
allows for package ids composed of multiple path components.
This removes the stacking of type parameters that occurs when invoking
trait methods, and fixes all places in the standard library that were
relying on it. It is somewhat awkward in places; I think we'll probably
want something like the `Foo::<for T>::new()` syntax.
Fixes for #8625 to prevent assigning to `&mut` in borrowed or aliasable locations. The old code was insufficient in that it failed to catch bizarre cases like `& &mut &mut`.
r? @pnkfelix
`UnsafeAtomicRcBox` → `UnsafeArc` (#7674), and `AtomicRcBoxData` → `ArcData` to reflect this.
Also, the inner pointer of `UnsafeArc` is now `*mut ArcData`, which avoids some transmutes to `~`: i.e. less chance of mistakes.
As for now, rekillable is an unsafe function, instead, it should behave
just like unkillable by encapsulating unsafe code within an unsafe
block.
This patch does that and removes unsafe blocks that were encapsulating
rekillable calls throughout rust's libs.
Fixes#8232
This means that fewer `transmute`s are required, so there is less
chance of a `transmute` not having the corresponding `forget`
(possibly leading to use-after-free, etc).
Now that new LLVM has landed, the debug info works on Windows as well. Most existing tests pass, except for the following four, which I left disabled for now:
lexical-scope-in-for-loop
lexical-scope-in-if
lexical-scope-in-match
lexical-scopes-in-block-expression
Also, fixed a small problem with the debug info test runner.
Beforehand, it was unclear whether rust was performing the "recommended set" of
optimizations provided by LLVM for code. This commit changes the way we run
passes to closely mirror that of clang, which in theory does it correctly. The
notable changes include:
* Passes are no longer explicitly added one by one. This would be difficult to
keep up with as LLVM changes and we don't guaranteed always know the best
order in which to run passes
* Passes are now managed by LLVM's PassManagerBuilder object. This is then used
to populate the various pass managers run.
* We now run both a FunctionPassManager and a module-wide PassManager. This is
what clang does, and I presume that we *may* see a speed boost from the
module-wide passes just having to do less work. I have no measured this.
* The codegen pass manager has been extracted to its own separate pass manager
to not get mixed up with the other passes
* All pass managers now include passes for target-specific data layout and
analysis passes
Some new features include:
* You can now print all passes being run with `-Z print-llvm-passes`
* When specifying passes via `--passes`, the passes are now appended to the
default list of passes instead of overwriting them.
* The output of `--passes list` is now generated by LLVM instead of maintaining
a list of passes ourselves
* Loop vectorization is turned on by default as an optimization pass and can be
disabled with `-Z no-vectorize-loops`
All of these "copies" of clang are based off their [source code](http://clang.llvm.org/doxygen/BackendUtil_8cpp_source.html) in case anyone is curious what my source is. I was hoping that this would fix#8665, but this does not help the performance issues found there. Hopefully i'll allow us to tweak passes or see what's going on to try to debug that problem.
Beforehand, it was unclear whether rust was performing the "recommended set" of
optimizations provided by LLVM for code. This commit changes the way we run
passes to closely mirror that of clang, which in theory does it correctly. The
notable changes include:
* Passes are no longer explicitly added one by one. This would be difficult to
keep up with as LLVM changes and we don't guaranteed always know the best
order in which to run passes
* Passes are now managed by LLVM's PassManagerBuilder object. This is then used
to populate the various pass managers run.
* We now run both a FunctionPassManager and a module-wide PassManager. This is
what clang does, and I presume that we *may* see a speed boost from the
module-wide passes just having to do less work. I have no measured this.
* The codegen pass manager has been extracted to its own separate pass manager
to not get mixed up with the other passes
* All pass managers now include passes for target-specific data layout and
analysis passes
Some new features include:
* You can now print all passes being run with `-Z print-llvm-passes`
* When specifying passes via `--passes`, the passes are now appended to the
default list of passes instead of overwriting them.
* The output of `--passes list` is now generated by LLVM instead of maintaining
a list of passes ourselves
* Loop vectorization is turned on by default as an optimization pass and can be
disabled with `-Z no-vectorize-loops`
