I tried a couple of different ways to squash this, and still don't think this is ideal, but I wanted to get it out for feedback.
Closes#5900Closes#9942
There are a few scenarios where the compiler tries to evaluate CastExprs without the corresponding types being available yet in the type context: https://github.com/mozilla/rust/issues/10618, https://github.com/mozilla/rust/issues/5900, https://github.com/mozilla/rust/issues/9942
This PR takes the approach of having eval_const_expr_partial's CastExpr arm fall back to a limited ast_ty_to_ty call that only checks for (a subset of) valid const types, when the direct type lookup fails. It's kind of hacky, so I understand if you don't want to take this as is. I'd need a little mentoring to get this into better shape, as figuring out the proper fix has been a little daunting. I'm also happy if someone else wants to pick this up and run with it.
This closes 5900 and 9942, but only moves the goalposts a little on 10618, which now falls over in a later phase of the compiler.
Previously, they were treated like ~[] and &[] (which can have length
0), but fixed length vectors are fixed length, i.e. we know at compile
time if it's possible to have length zero (which is only for [T, .. 0]).
Fixes#11659.
Previously, they were treated like ~[] and &[] (which can have length
0), but fixed length vectors are fixed length, i.e. we know at compile
time if it's possible to have length zero (which is only for [T, .. 0]).
Fixes#11659.
NodeIds are sequential integers starting at zero, so we can achieve some
memory savings by just storing the items all in a line in a vector.
The occupancy for typical crates seems to be 75-80%, so we're already
more efficient than a HashMap (maximum occupancy 75%), not even counting
the extra book-keeping that HashMap does.
Unsuffixed literals like 1 and 1.1, and free type parameters sometimes
have to be printed in error messages, which ended up with <V0>, <VI0>
and <VF0>. This change puts the words "generic" and "integer"/"float"
into the message so it's not a completely black box.
If it's a trait method, this checks the stability attribute of the
method inside the trait definition. Otherwise, it checks the method
implementation itself.
Close#8961.
If it's a trait method, this checks the stability attribute of the
method inside the trait definition. Otherwise, it checks the method
implementation itself.
### Rationale
There is no reason to support more than 2³² nodes or names at this moment, as compiling something that big (even without considering the quadratic space usage of some analysis passes) would take at least **64GB**.
Meanwhile, some can't (or barely can) compile rustc because it requires almost **1.5GB**.
### Potential problems
Can someone confirm this doesn't affect metadata (de)serialization? I can't tell myself, I know nothing about it.
### Results
Some structures have a size reduction of 25% to 50%: [before](https://gist.github.com/luqmana/3a82a51fa9c86d9191fa) - [after](https://gist.github.com/eddyb/5a75f8973d3d8018afd3).
Sadly, there isn't a massive change in the memory used for compiling stage2 librustc (it doesn't go over **1.4GB** as [before](http://huonw.github.io/isrustfastyet/mem/), but I can barely see the difference).
However, my own testcase (previously peaking at **1.6GB** in typeck) shows a reduction of **200**-**400MB**.
Bare functions are another example of a scalar but non-numeric
type (like char) that should be handled separately in casts.
This disallows expressions like `0 as extern "Rust" fn() -> int;`.
It might be advantageous to allow casts between bare functions
and raw pointers in unsafe code in the future, to pass function
pointers between Rust and C.
Closes#8728
than the current ones, which were very fine-grained. Also, cleanly distinguish
when properties must be found in *owned* types vs *reachable* types.
Fixes#10157Fixes#10278
This isn't quite as fancy as the struct in #9913, but I'm not sure we should be exposing crate names/hashes of the types. That being said, it'd be pretty easy to extend this (the deterministic hashing regardless of what crate you're in was the hard part).
Cleaned up the source in a few places
Renamed `map_move` to `map`, removed other `map` methods
Added `as_ref` and `as_mut` adapters to `Result`
Added `fmt::Default` impl
Allows an enum with a discriminant to use any of the primitive integer types to store it. By default the smallest usable type is chosen, but this can be overridden with an attribute: `#[repr(int)]` etc., or `#[repr(C)]` to match the target's C ABI for the equivalent C enum.
Also adds a lint pass for using non-FFI safe enums in extern declarations, checks that specified discriminants can be stored in the specified type if any, and fixes assorted code that was assuming int.
Previously an ExprLit was created *per byte* causing a huge increase in memory
bloat. This adds a new `lit_binary` to contain a literal of binary data, which
is currently only used by the include_bin! syntax extension. This massively
speeds up compilation times of the shootout-k-nucleotide-pipes test
before:
time: 469s
memory: 6GB
assertion failure in LLVM (section too large)
after:
time: 2.50s
memory: 124MB
Closes#2598
Bare functions are another example of a scalar but non-numeric
type (like char) that should be handled separately in casts.
This disallows expressions like `0 as extern "Rust" fn() -> int;`.
It might be advantageous to allow casts between bare functions
and raw pointers in unsafe code in the future, to pass function
pointers between Rust and C.
Closes#8728
For the benefit of the pretty printer we want to keep track of how
string literals in the ast were originally represented in the source
code.
This commit changes parser functions so they don't extract strings from
the token stream without at least also returning what style of string
literal it was. This is stored in the resulting ast node for string
literals, obviously, for the package id in `extern mod = r"package id"`
view items, for the inline asm in `asm!()` invocations.
For `asm!()`'s other arguments or for `extern "Rust" fn()` items, I just
the style of string, because it seemed disproportionally cumbersome to
thread that information through the string processing that happens with
those string literals, given the limited advantage raw string literals
would provide in these positions.
The other syntax extensions don't seem to store passed string literals
in the ast, so they also discard the style of strings they parse.
According to http://huonw.github.io/isrustfastyet/mem/#012f909, the "const
marking" pass generates about 400MB of extra memory during compilation. It
appears that this is due to two different factors:
1. There is a `ccache` map in the ty::ctxt which is only ever used in this
pass, so this commit moves the map out of the ty::ctxt struct and into
just this pass's visitor. This turned out to not benefit that much in
memory (as indicated by http://i.imgur.com/Eo4iOzK.png), but it's helpful
to do nonetheless.
2. During const_eval, there are a lot of lookups into decoding inlined items
from external crates. There is no caching involved here, so the same
static or variant could be re-translated many times. After adding
separate caches for variants and statics, the memory peak of compiling
rustc decreased by 200MB (as evident by http://i.imgur.com/ULAUMtq.png)
The culmination of this is basically a slight reorganization of a caching map
for the const_eval pass along with a 200MB decrease in peak memory usage when
compiling librustc.
It is simply defined as `f64` across every platform right now.
A use case hasn't been presented for a `float` type defined as the
highest precision floating point type implemented in hardware on the
platform. Performance-wise, using the smallest precision correct for the
use case greatly saves on cache space and allows for fitting more
numbers into SSE/AVX registers.
If there was a use case, this could be implemented as simply a type
alias or a struct thanks to `#[cfg(...)]`.
Closes#6592
The mailing list thread, for reference:
https://mail.mozilla.org/pipermail/rust-dev/2013-July/004632.html
This is mostly for consistency, as you can now compare raw pointers in
constant expressions or without the standard library.
It also reduces the number of `ptrtoint` instructions in the IR, making
tracking down culprits of what's usually an anti-pattern easier.
This is mostly for consistency, as you can now compare raw pointers in
constant expressions or without the standard library.
It also reduces the number of `ptrtoint` instructions in the IR, making
tracking down culprits of what's usually an anti-pattern easier.
Also redefine all of the standard logging macros to use more rust code instead
of custom LLVM translation code. This makes them a bit easier to understand, but
also more flexibile for future types of logging.
Additionally, this commit removes the LogType language item in preparation for
changing how logging is performed.
Also redefine all of the standard logging macros to use more rust code instead
of custom LLVM translation code. This makes them a bit easier to understand, but
also more flexibile for future types of logging.
Additionally, this commit removes the LogType language item in preparation for
changing how logging is performed.
There are 6 new compiler recognised attributes: deprecated, experimental,
unstable, stable, frozen, locked (these levels are taken directly from
Node's "stability index"[1]). These indicate the stability of the
item to which they are attached; e.g. `#[deprecated] fn foo() { .. }`
says that `foo` is deprecated.
This comes with 3 lints for the first 3 levels (with matching names) that
will detect the use of items marked with them (the `unstable` lint
includes items with no stability attribute). The attributes can be given
a short text note that will be displayed by the lint. An example:
#[warn(unstable)]; // `allow` by default
#[deprecated="use `bar`"]
fn foo() { }
#[stable]
fn bar() { }
fn baz() { }
fn main() {
foo(); // "warning: use of deprecated item: use `bar`"
bar(); // all fine
baz(); // "warning: use of unmarked item"
}
The lints currently only check the "edges" of the AST: i.e. functions,
methods[2], structs and enum variants. Any stability attributes on modules,
enums, traits and impls are not checked.
[1]: http://nodejs.org/api/documentation.html
[2]: the method check is currently incorrect and doesn't work.
