Commit Graph

174 Commits

Author SHA1 Message Date
Daniel Micay
22b6f7481f document why attributes are set on CallInst 2013-09-16 13:44:04 -04:00
Daniel Micay
b2eb1c01a4 add sret + noalias to the out pointer parameter
This brings Rust in line with how `clang` handles return pointers.

Example:

    pub fn bar() -> [uint, .. 8] {
        let a = [0, .. 8];
        a
    }

Before:

    ; Function Attrs: nounwind uwtable
    define void @_ZN3bar17ha4635c6f704bfa334v0.0E([8 x i64]* nocapture, { i64, %tydesc*, i8*, i8*, i8 }* nocapture readnone) #1 {
    "function top level":
      %a = alloca [8 x i64], align 8
      %2 = bitcast [8 x i64]* %a to i8*
      call void @llvm.memset.p0i8.i64(i8* %2, i8 0, i64 64, i32 8, i1 false)
      %3 = bitcast [8 x i64]* %0 to i8*
      call void @llvm.memcpy.p0i8.p0i8.i64(i8* %3, i8* %2, i64 64, i32 8, i1 false)
      ret void
    }

After:

    ; Function Attrs: nounwind uwtable
    define void @_ZN3bar17ha4635c6f704bfa334v0.0E([8 x i64]* noalias nocapture sret, { i64, %tydesc*, i8*, i8*, i8 }* nocapture readnone) #1 {
    "function top level":
      %2 = bitcast [8 x i64]* %0 to i8*
      call void @llvm.memset.p0i8.i64(i8* %2, i8 0, i64 64, i32 8, i1 false)
      ret void
    }

Closes #9072
Closes #7298
Closes #9154
2013-09-16 13:44:04 -04:00
Daniel Micay
6d0a847c3a teach Call/CallWithConv to set attributes 2013-09-15 23:34:11 -04:00
SiegeLord
8c5402fd36 Properly encode/decode structural variants. 2013-09-11 14:49:09 -04:00
Marvin Löbel
7419085337 Modernized a few more types in syntax::ast 2013-09-03 14:45:06 +02:00
Patrick Walton
6c37e3b7f8 librustc: Implement basic lazy implementation loading.
This is only for implementations defined in the same crate as the trait they
implement.
2013-08-27 18:47:59 -07:00
Patrick Walton
8693943676 librustc: Ensure that type parameters are in the right positions in paths.
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.
2013-08-27 18:47:57 -07:00
bors
9feaf1d023 auto merge of #8594 : bytewiseand/rust/static-fn-ptr, r=pcwalton
Fixes #8588
2013-08-21 15:51:34 -07:00
bors
2246d56e71 auto merge of #8619 : pnkfelix/rust/fsk-visitor-vpar-defaults-step3, r=nmatsakis
Follow up to #8539 (step 2 of 5).

(See  #8527, which was step 1 of 5, for the full outline.)

Part of #7081.
2013-08-19 09:12:03 -07:00
Niko Matsakis
303f650ecf Issue #3678: Remove wrappers and call foreign functions directly 2013-08-19 07:13:15 -04:00
Andreas Martens
a1066130c0 Fix ICE when calling static and static function pointers
Fixes #8588
2013-08-18 20:21:23 +02:00
Felix S. Klock II
23ce08a4cd port callee.rs from oldvisit to <V:Visitor> trait API. 2013-08-15 04:09:19 +02:00
Niko Matsakis
006c6b6be4 trans: Rely on new AutoBorrowObj adjustment to match up object receivers
Note: some portions of this commit written by @Sodel-the-Vociferous
(Daniel Ralston)
2013-08-11 14:01:19 -04:00
bors
8495ee52b2 auto merge of #8262 : dotdash/rust/no_rval_copies, r=pcwalton
rvalues aren't going to be used anywhere but as the argument, so
there's no point in copying them. LLVM used to eliminate the copy
later, but why bother emitting it in the first place?
2013-08-04 10:55:53 -07:00
Daniel Micay
e7bb33aed8 rm obsolete for support from the compiler 2013-08-04 00:39:48 -04:00
Daniel Micay
1008945528 remove obsolete foreach keyword
this has been replaced by `for`
2013-08-03 22:48:02 -04:00
Björn Steinbrink
a51e3e46ef trans_arg_expr: Omit extra copies for rvalues
rvalues aren't going to be used anywhere but as the argument, so
there's no point in copying them. LLVM used to eliminate the copy
later, but why bother emitting it in the first place?
2013-08-03 15:03:58 +02:00
Patrick Walton
887c656970 librustc: Introduce a new visitor type based on traits and port syntax to it.
This is preparation for removing `@fn`.

This does *not* use default methods yet, because I don't know
whether they work. If they do, a forthcoming PR will use them.

This also changes the precedence of `as`.
2013-08-02 21:57:58 -07:00
Daniel Micay
1fc4db2d08 migrate many for loops to foreach 2013-08-01 05:34:55 -04:00
Michael Woerister
8a329770b6 New naming convention for ast::{node_id, local_crate, crate_node_id, blk_check_mode, ty_field, ty_method} 2013-07-29 16:16:48 +02:00
bors
baa649ede6 auto merge of #8027 : nikomatsakis/rust/issue-4846-multiple-lifetime-parameters-1, r=pcwalton
Small step towards #4846. r? @msullivan
2013-07-25 07:37:45 -07:00
bors
906264b50f auto merge of #8015 : msullivan/rust/default-methods, r=nikomatsakis
Lots of changes to vtable resolution, handling of super/self method calls in default methods. Fix a lot of trait inheritance bugs.

r? @nikomatsakis
2013-07-25 03:07:44 -07:00
Niko Matsakis
dc38e1616a Generalize the ty::substs struct so that it can represent
multiple lifetime parameters, and not just one. Also add an option
for erasing lifetimes, which makes trans code somewhat simpler
and cleaner.
2013-07-24 16:52:57 -04:00
Birunthan Mohanathas
d047cf1ec6 Change 'print(fmt!(...))' to printf!/printfln! in src/lib* 2013-07-24 09:45:20 -04:00
Michael Sullivan
79f8a7fee5 Properly track and export information about vtables for impls in metadata.
Partially rework how vtables are handled in default method calls.
Closes #7460.
2013-07-23 17:06:33 -07:00
Michael Sullivan
89c4af0ea9 Fix a default method polymorphism bug. 2013-07-23 17:06:32 -07:00
Michael Woerister
3315edfae7 Adapted trans::common::{block, fn_ctxt, scope_info} to new naming convention. 2013-07-23 15:38:55 +02:00
Michael Woerister
4bd1424622 Ast spanned<T> refactoring, renaming: crate, local, blk, crate_num, crate_cfg.
`crate => Crate`
`local => Local`
`blk => Block`
`crate_num => CrateNum`
`crate_cfg => CrateConfig`

Also, Crate and Local are not wrapped in spanned<T> anymore.
2013-07-22 15:35:28 +02:00
Michael Sullivan
585e283769 Add provided method information to ty::Method. Get rid of ProvidedMethodSource. 2013-07-18 13:56:14 -07:00
Patrick Walton
99b33f7219 librustc: Remove all uses of "copy". 2013-07-17 14:57:51 -07:00
Patrick Walton
b4e674f6e6 librustc: Add a lint mode for unnecessary copy and remove a bunch of them. 2013-07-17 14:56:42 -07:00
Alex Crichton
242606c793 Clean up various warnings throughout the codebase 2013-07-14 09:29:12 -07:00
Björn Steinbrink
dcd5d14e6c Avoid return blocks that have only a single predecessor
Currently, we always create a dedicated "return" basic block, but when
there's only a single predecessor for that block, it can be merged with
that predecessor. We can achieve that merge by only creating the return
block on demand, avoiding its creation when its not required.

Reduces the pre-optimization size of librustc.ll created with --passes ""
by about 90k lines which equals about 4%.
2013-07-13 13:33:48 +02:00
Michael Sullivan
19446418bc Drop a now unnecessary argument from maybe_instantiate_inline. 2013-07-11 15:51:10 -07:00
Michael Sullivan
2b89b437fb Normalize self types for monomorphization. 2013-07-11 15:51:09 -07:00
Brendan Cully
e6e4f52bcf remove unused imports 2013-07-10 22:08:50 -07:00
Niko Matsakis
979d3a54f9 Correct merge failures 2013-07-08 13:55:11 -04:00
Niko Matsakis
2d3262ca7b Update trans to use type to decide when to move, not the moves table (simpler
for cases where it's hard to decide what id to use for the lookup); modify
irrefutable bindings code to move or copy depending on the type, rather than
threading through a flag. Also updates how local variables and arguments are
registered. These changes were hard to isolate.
2013-07-08 13:55:10 -04:00
Niko Matsakis
41efcdf299 Make all allocas named so we can see where they originate
in the generated LLVM code.
2013-07-08 13:55:10 -04:00
Björn Steinbrink
00ba8b3ac0 Improve handling of immediate return values
We currently still handle immediate return values a lot like
non-immediate ones. We provide a slot for them and store them into
memory, often just to immediately load them again. To improve this
situation, trans_call_inner has to return a Result which contains the
immediate return value.

Also, it also needs to accept "No destination" in addition to just
SaveIn and Ignore. Since "No destination" isn't something that fits
well into the Dest type, I've chosen to simply use Option<Dest>
instead, paired with an assertion that checks that "None" is only
allowed for immediate return values.
2013-07-08 13:34:13 +02:00
Björn Steinbrink
4a485f8cec Avoid unused allocas for immediate return values
There's no need to allocate a return slot for anykind of immediate
return value, not just not for nils. Also, when the return value is
ignored, we only have to copy it to a temporary alloca if it's actually
required to call drop_ty on it.
2013-07-08 13:17:46 +02:00
Seo Sanghyeon
fd19289108 Classify newtype structs S(T) as immediates if T is an immediate 2013-07-01 13:02:14 +09:00
bors
6fcd8bf567 auto merge of #7468 : cmr/rust/great_renaming, r=pcwalton 2013-06-30 01:19:38 -07:00
Björn Steinbrink
765a2901d5 Avoid double indirection for the "self" arg in methods
Currently we pass all "self" arguments by reference, for the pointer
variants this means that we end up with double indirection which causes
a unnecessary performance hit.

The fix itself is pretty straight-forward and just means that "self"
needs to be handled like any other argument, except for by-value "self"
which still needs to be passed by reference. This is because
non-pointer types can't just be stuffed into the environment slot which
is used to pass "self".

What made things tricky is that there was also a bug in the typechecker
where the method map entries are created. For type impls, that stored
the base type instead of the actual self-type in the method map, e.g.
Foo instead of &Foo for &self. That worked with pass-by-reference, but
fails with pass-by-value which needs the real type.

Code that makes use of methods seems to be about 10% faster with this
change. Also, build times are reduced by about 4%.

Fixes #4355, #4402, #5280, #4406 and #7285
2013-06-29 19:27:40 +02:00
Björn Steinbrink
ba922a4027 Fix handling of temp cleanups for the "self" argument
The code that tried to revoke the cleanup for the self argument tried
to use "llself" to do so, but the cleanup might actually be registered
with a different ValueRef due to e.g. casting. Currently, this is
worked around by early revocation of the cleanup for self in
trans_self_arg.

To handle this correctly, we have to put the ValueRef for the cleanup
into the MethodData, so trans_call_inner can use it to revoke the
cleanup when it's actually supposed to.
2013-06-29 18:02:21 +02:00
Corey Richardson
71b1c6ab60 Warning cleanup 2013-06-29 11:20:04 -04:00
Corey Richardson
95d22a3c81 Fix merge fallout 2013-06-29 11:20:04 -04:00
Corey Richardson
1662bd371c Great renaming: propagate throughout the rest of the codebase 2013-06-29 11:20:02 -04:00
Björn Steinbrink
2a40c5db46 Avoid extra casts for "self" arguments
"self" is always passed as an opaque box, so there's no point in using
the concrete self type when translating the argument. All it does it
causing the value to be casted back to an opaque box right away.
2013-06-29 17:08:20 +02:00
Björn Steinbrink
08a5278bb5 Add missing argument to a debug! call in trans_arg_expr 2013-06-29 15:59:33 +02:00