aboutsummaryrefslogtreecommitdiff
path: root/src/comp/middle/trans.rs
Commit message (Collapse)AuthorAgeFilesLines
...
* Join the results of alt arms so alt can be used as an expression.Brian Anderson2011-04-021-12/+18
| | | | Un-XFAIL expr-alt.rs.
* Make join_results branch into the join block even when only one edge is live.Brian Anderson2011-04-021-6/+0
| | | | | | I don't don't totally understand the implications of this but it makes the behavior consistent for all live edges, which is going to make joining the arms of an alt expression work correctly.
* Fail on non-exhaustive alt matchesBrian Anderson2011-04-021-3/+3
|
* Terminate fail blocks by making their ends unreachable.Brian Anderson2011-04-021-3/+3
| | | | This will make implementing default alt arms simpler.
* Support for passing native types through metadata/creader.Graydon Hoare2011-04-011-1/+3
|
* Guh. Actually commit the staged undo-of-indentation. Forgotten git staging ↵Graydon Hoare2011-04-011-56/+56
| | | | wins again.
* Switch private linkage to local, re-indent trans.rs.Graydon Hoare2011-04-011-70/+70
|
* rustc: Get tag variants from the crate metadataPatrick Walton2011-04-011-1/+4
|
* rustc: Have tag_variants() return a new variant_info structure instead of an ↵Patrick Walton2011-04-011-37/+40
| | | | AST node
* Started adding support for typestate checking.Tim Chevalier2011-04-011-11/+12
| | | | | | | | | | | | | | | I added a new field to the ast "ann" type for typestate information. Currently, the field contains a record of a precondition bit vector and postcondition vector, but I tried to structure things so as to make it easy to change the representation of the typestate annotation type. I also had to add annotations to some syntactic forms that didn't have them before (fail, ret, be...), with all the boilerplate changes that that would imply. The main call to the typestate_check entry point is commented out and the actual pre-postcondition algorithm only has a few cases implemented, though the overall AST traversal is there. The rest of the typestate algorithm isn't implemented yet.
* rustc: Remove useless call to tag_variant_with_id()Patrick Walton2011-04-011-1/+0
|
* Cleanup typos and newlinesBrian Anderson2011-04-011-1/+1
|
* Initialize the alloca used to retrieve boxed block results to null.Brian Anderson2011-03-311-11/+19
| | | | | | | | This allows blocks to be used in conditional constructs where the block may not ever execute: the drop glue will notice that it was never used and ignore it. Also, beef up the comments.
* Only use allocas to hold expression results for boxed typesBrian Anderson2011-03-311-18/+3
|
* Add more comments to trans_blockBrian Anderson2011-03-311-1/+2
|
* Modify trans.rs to allow blocks to result in boxed valuesBrian Anderson2011-03-311-0/+59
|
* rustc: Add a type annotation to tag itemsPatrick Walton2011-03-311-7/+7
|
* rustc: Make tag_ty_params() and substitute_ty_params() take def ids instead ↵Patrick Walton2011-03-311-7/+3
| | | | of ty_params, and to check in external crates
* More stuff to go with the new expr_call_self AST nodeLindsey Kuper2011-03-311-0/+68
|
* rustc: Use the scaled index, not the raw index, if a vector has generic ↵Patrick Walton2011-03-311-2/+12
| | | | size. lib-vec.rs works now.
* Rename incr_all_refcnts to take_ty, to match drop_tyBrian Anderson2011-03-311-4/+4
|
* Restore scope block contexts to the translation of expr_block.Brian Anderson2011-03-311-2/+20
| | | | | I removed this previously but that was wrong because it caused block expressions to not create a new scope.
* Improve line comment positioningMarijn Haverbeke2011-03-311-23/+24
| | | | This involved making ast.variant spanned.
* rustc: Use the type cache to construct external item referencesPatrick Walton2011-03-301-19/+38
|
* rustc: Thread an item-to-type mapping throughout the typechecking and ↵Patrick Walton2011-03-301-21/+19
| | | | translation phases
* Ensure task stacks start out with a 16-byte aligned entry frame. Should make ↵Graydon Hoare2011-03-301-1/+2
| | | | OSX behave a bit better.
* rustc: Add all items to the item table during translation. lib-vec now ↵Patrick Walton2011-03-291-1/+16
| | | | compiles, but asserts and then segfaults.
* Rename trans_native to trans_native_call, for clarity.Graydon Hoare2011-03-291-10/+10
|
* Teach trans to emit undefined references to 'use'd symbols. Can compile and ↵Graydon Hoare2011-03-291-0/+21
| | | | run a simple 'use std;' example now.
* Add partial support for if expressions resulting in generic typesBrian Anderson2011-03-281-4/+7
|
* Handle structural types as the result of if expressionsBrian Anderson2011-03-281-0/+3
|
* Teach rustc to use the result of if expressionsBrian Anderson2011-03-281-4/+13
|
* Move all allocas to dedicated basic block at top of function, to keep frames ↵Graydon Hoare2011-03-281-56/+141
| | | | finite. Un-comment next size in nbody.rs.
* Fix get_tydesc to alloca enough room for root; remove hack that was ↵Graydon Hoare2011-03-281-8/+3
| | | | attempting to paper over it. Fix from espindola.
* rustc: Add support for calling LLVM intrinsics as native functionsPatrick Walton2011-03-281-8/+34
|
* Add support for break and cont to rustcMarijn Haverbeke2011-03-271-9/+72
| | | | | Testing proper cleanup is hampered by https://github.com/graydon/rust/issues/293
* Twiddle visibility, start exposing only type-mangled names (64 bit truncated ↵Graydon Hoare2011-03-261-17/+21
| | | | sha1 prefixes).
* Start refactoring name-mangling in trans.Graydon Hoare2011-03-261-43/+90
|
* Add get_extern_const, factor get_extern into get_extern_fn and ↵Graydon Hoare2011-03-251-6/+23
| | | | get_simple_extern_fn.
* Trans nomenclature tidy-up: upcall vs. native vs. extern.Graydon Hoare2011-03-251-35/+54
|
* Another go at changing compile-command, this time using RBUILD env var.Graydon Hoare2011-03-251-1/+1
|
* Revert "Bulk-edit compile commands in emacs chatter to point to assumed ↵Graydon Hoare2011-03-251-1/+1
| | | | | | build/ dir off src root." This reverts commit 846f2e2ba994268725f38c36fa12f1a09f21615c.
* Bulk-edit compile commands in emacs chatter to point to assumed build/ dir ↵Graydon Hoare2011-03-251-1/+1
| | | | off src root.
* Implement local declarations with receive. Un-XFAIL decl-with-recv.rs.Brian Anderson2011-03-251-7/+23
|
* Refactor ast.local to make room for initialization via recvBrian Anderson2011-03-251-2/+2
|
* Revert "Change the type of the second argument of upcalls to be a %task*.", ↵Graydon Hoare2011-03-231-13/+13
| | | | | | breaks win32 build. This reverts commit dcf122ede03af71d8a1d892a077a9e2435b64ef3.
* rustc: Write out crate metadataPatrick Walton2011-03-231-1/+11
|
* Further support for floating-point. Literals with exponents workTim Chevalier2011-03-221-14/+40
| | | | | | and literals with the 'f32' or 'f64' suffixes work as well. In addition, logging things with the f32 or f64 type works. (float is still assumed to be a synonym for f64).
* Change the type of the second argument of upcalls to be a %task*.Rafael Ávila de Espíndola2011-03-221-13/+13
|
* Revert "Remove usages of case(_) { fail; } since the compiler does this ↵Patrick Walton2011-03-221-0/+3
| | | | | | automatically". When we have exhaustiveness checking, "case(_) { fail; }" will be useful to silence warnings. This reverts commit 92a716d862d92d3cc52a400457d2c3900d0c57a2.