Eric Holk
c3b98cabe1
Removed pretty much all the vector+ from core (issue #2719 )
2012-06-27 15:22:06 -07:00
Graydon Hoare
e2205659e3
Fix tabSize doc build breakage, also remove the stray tab that got into docs.
2012-06-27 14:37:43 -07:00
Patrick Walton
c1157161d9
rustc: Don't give variables the same name as enums in trans
...
Conflicts:
src/rustc/middle/trans/alt.rs
src/rustc/middle/trans/base.rs
src/rustc/middle/trans/closure.rs
src/rustc/middle/trans/impl.rs
src/rustc/middle/trans/uniq.rs
2012-06-27 14:11:02 -07:00
Patrick Walton
6cf5910de2
rustc: Don't allow multiple candidate methods from impls with the same def ID
2012-06-27 14:11:02 -07:00
Graydon Hoare
576f59a79b
Change 'upcall fail' to 'task failed at', close #2678 .
2012-06-27 13:30:49 -07:00
Graydon Hoare
d3c6119a7a
More keyword paring and migration in runtime, docs, code modes.
2012-06-27 13:30:49 -07:00
Lindsey Kuper
f0565be49a
Better error message instead of bare "fail" in infer
2012-06-27 12:37:50 -07:00
Eric Holk
133fdc1148
Remove unnecessary bounds checks in vec::push_all (issue #2719 )
...
Don't needlessly drop closures (issue #2603 )
2012-06-27 11:32:22 -07:00
Eric Holk
404afcbb41
Mark exclusive arc-related functions as unsafe. Fixes #2727 .
2012-06-27 10:24:51 -07:00
Graydon Hoare
697f1e38d6
Change 'native' and 'crust' to 'extern'.
...
This comes with a terminology change. All linkage-symbols are 'extern'
now, including rust syms in other crates. Some extern ABIs are
merely "foreign". The term "native" is retired, not clear/useful.
What was "crust" is now "extern" applied to a _definition_. This
is a bit of an overloading, but should be unambiguous: it means
that the definition should be made available to some non-rust ABI.
2012-06-26 16:18:37 -07:00
Patrick Walton
999b567e2a
rustdoc: Work around some more metadata infelicities
2012-06-26 14:39:49 -07:00
Patrick Walton
d4672a078e
libstd: Work around some metadata infelicities that are causing problems in resolve3, and remove an unused glob
...
Conflicts:
src/libstd/net_tcp.rs
src/libstd/par.rs
2012-06-26 14:37:29 -07:00
Eric Holk
ccd2394ff9
perf: improved the core-std benchmark
2012-06-26 14:08:16 -07:00
Eric Holk
a08281616f
More perf tweaks (issue #2719 )
2012-06-26 14:08:16 -07:00
Michael Sullivan
51468b65a4
Properly cleanup slice literals. Closes #2705 .
2012-06-26 14:05:43 -07:00
Michael Sullivan
f17ca3f73e
Some box cleanup that doesn't break the build.
2012-06-26 13:58:21 -07:00
Patrick Walton
6413421c8c
core: make_a_sandwitch -> make_a_sandwich
2012-06-26 13:55:33 -07:00
Patrick Walton
481267299d
rustc: Make trans no longer insist that there be exactly one impl scope per module
...
Conflicts:
src/rustc/middle/trans/base.rs
2012-06-26 13:45:18 -07:00
Brian Anderson
1228c55ef0
Register snapshots
2012-06-26 13:42:35 -07:00
Michael Sullivan
f85fbcb67f
Revert "Clean up a bunch of box related code."
...
This reverts commit bacf9e9887
2012-06-26 12:58:58 -07:00
Tim Chevalier
6141f5ce5a
Fix shape::cmp::walk_res2
2012-06-26 11:14:39 -07:00
Tim Chevalier
91e44ea056
Be a little more clever about calculating sizes for class types
...
We could try to calculate the size of a partially-converted class type,
and run into an LLVM error because we were trying to calculate the size
of a named struct whose fields hadn't yet be filled in. The issue can be
dodged by extending simplify_type to convert classes to isomorphic
structural records, just for the purposes of size calculations.
(for real this time) Closes #2718
2012-06-26 11:14:39 -07:00
Tim Chevalier
cf69604551
Incorporate class fields into recursive-type check
...
Noticed while investigating issue 2718 that the typechecker allowed
some non-instantiable types involving classes. This wasn't the root of
2718, but fixed it anyway.
2012-06-26 11:14:39 -07:00
Tim Chevalier
d513d9893e
Typos in a comment and an error message
2012-06-26 11:14:38 -07:00
Michael Sullivan
bacf9e9887
Clean up a bunch of box related code.
2012-06-26 10:03:10 -07:00
Eric Holk
b9d3ad0736
Getting rid of lots more vector +=. (issue #2719 )
2012-06-26 00:39:18 -07:00
Brian Anderson
007b9d9acd
Update combine-tests.py for new vec syntax
2012-06-25 21:38:27 -07:00
Michael Sullivan
15b60ac56b
Comments only: fix some comments that got spurious /~s
2012-06-25 20:35:32 -07:00
Michael Sullivan
329eca6044
Make vectors uglier ([]/~). Sorry. Should be temporary. Closes #2725 .
2012-06-25 20:00:46 -07:00
Michael Sullivan
c087aaf56b
When we cons up vector asts, generate evecs.
2012-06-25 20:00:39 -07:00
Michael Sullivan
a71a49faa8
Split deprecated str and vec warnings into two flags, enable old_vecs by default.
2012-06-25 19:29:27 -07:00
Michael Sullivan
1c5e2cacef
Comment out the *right* part of the test failing due to issue #2724 .
2012-06-25 19:28:27 -07:00
Michael Sullivan
fb018fa50b
Comment out a failing part of a test; this is Issue #2724 .
2012-06-25 19:00:35 -07:00
Paul Stansifer
4df2654f82
Make it possible to make built-in tt syntax extensions
2012-06-25 18:01:37 -07:00
Paul Stansifer
4f104954a6
parsing for the macro system
2012-06-25 18:01:37 -07:00
Patrick Walton
650dfe58a3
test: Make zip-same-length no longer rely on resolve not handling shadowed modules properly
2012-06-25 17:55:24 -07:00
Brian Anderson
c904927f73
Merge pull request #2722 from bstrie/py2
...
Fail to configure without an LLVM-friendly Python
2012-06-25 17:54:52 -07:00
Brian Anderson
c2751c5640
rustc: Make 'do' work with effects the same as 'for'
2012-06-25 17:46:06 -07:00
Tim Chevalier
8ed9701591
Remove test case that doesn't work on its own
...
I don't actually know how to test --ls
2012-06-25 17:26:26 -07:00
Tim Chevalier
f1acc69a2a
Add class fields to the global index
...
Closes #2192
2012-06-25 17:25:54 -07:00
Ben Striegel
6322eda35c
Fail to configure without an LLVM-friendly Python
...
This addresses issue #2720 . According to LLVM's documentation, it requires a
version of Python between 2.4 and 2.7. Without the proper version, LLVM fails
to build with cryptic errors. Prior to this commit, the configure script
checked for the `python` command in the environment, but didn't actually check
the version, which can cause problems e.g. on Linux distros where the default
is Python 3. Now the configure script always prefers to select a more specific
version of Python when available, in the order `python2.7` > `python2.6` >
`python2` > `python`, and will always check to ensure that the interpreter's
version is in the correct range.
2012-06-25 20:18:09 -04:00
Eric Holk
b837f37d40
vec::slice is faster now (Issue #2719 )
2012-06-25 17:08:06 -07:00
Eric Holk
b19c98ea9a
Some perf fixes, although vec::slice is still too slow (Issue #2719 )
2012-06-25 16:38:12 -07:00
Brian Anderson
7adad4c6cb
rt: Fix exports
2012-06-25 16:30:42 -07:00
Brian Anderson
d29e8fc059
rt: Remove upcall_malloc/upcall_exchange_malloc. Issue #2681
2012-06-25 16:01:47 -07:00
Brian Anderson
8910b2ff7c
build: Remove bogus -fno-strict-aliasing flags. Issue #2701
2012-06-25 15:41:03 -07:00
Brian Anderson
3d0826b5fc
rt: Remove lock_free_queue. Unused. Issue #2701
2012-06-25 15:39:59 -07:00
Patrick Walton
216105fc55
test: Remove the import-glob-circular test since the name resolution semantics it depends on are probably undecidable
2012-06-25 15:30:38 -07:00
Patrick Walton
9386c2db33
test: Swap the order of two imports in import7 to eliminate another cycle
2012-06-25 15:14:28 -07:00
Patrick Walton
d8410c7578
test: Swap the order of two imports in import6 to eliminate a cycle
2012-06-25 15:12:50 -07:00
