Commit Graph

6 Commits

Author SHA1 Message Date
John McCall
a1b469ed2f ExplosionKind -> ResilienceExpansion. NFC.
Swift SVN r12364
2014-01-16 00:25:29 +00:00
Joe Groff
990212b1c6 SILGen: When applying the result function of a function, do not reabstract to the original callee's abstraction pattern.
We already reabstracted the result function to the substituted function type. Quick fix for <rdar://problem/15598596>.

Swift SVN r12307
2014-01-15 01:20:14 +00:00
Joe Groff
7331eddfcc SIL: Handle dependent types in AbstractionPattern::isOpaque.
To do this with complete accuracy requires mapping the dependent types into a context, but isOpaque is currently only used as a sanity check in assertions, and a tuple being matched to a class-constrained archetype is an unlikely enough occurrence that it's probably OK to accept this inaccuracy instead of trying to push generic contexts all the way down into isOpaque's clients.

Swift SVN r11824
2014-01-02 19:04:00 +00:00
Chris Lattner
1cda63d0b3 sil abstraction patterns never see @lvalue types, only @inout types.
Swift SVN r11797
2014-01-01 20:24:34 +00:00
Chris Lattner
9ae289de46 Drive the semantic wedge harder into lvalues. Now, instead of having one LValueType
with qualifiers on it, we have two distinct types:
 - LValueType(T) aka @lvalue T, which is used for mutable values on the LHS of an
   assignment in the typechecker.
 - InOutType(T) aka @inout T, which is used for @inout arguments, and the implicit
   @inout self argument of mutable methods on value types.  This type is also used
   at the SIL level for address types.

While I detangled a number of cases that were checking for LValueType (without checking
qualifiers) and only meant @inout or @lvalue, there is more to be done here.  Notably,
getRValueType() still strips @inout, which is totally and unbearably wrong.



Swift SVN r11727
2013-12-29 22:23:11 +00:00
John McCall
4bba9b38f8 Make several new interfaces traffic in AbstractionPatterns.
Swift SVN r10621
2013-11-21 02:19:46 +00:00