[PATCH] [Sema] Re-land "Make FunctionType's TSI use unadjusted argument types"

[Reid Kleckner]

For a Type held by an AST node, I completely agree, and I'll try to address
that.  For a type held by a TypeSourceInfo, I disagree, the
FunctionProtoType should really hold the undecayed parameter types.

The problem is that we have lots of call sites where Sema pulls the type
off of TSI and builds it into the AST.  I'll see if I can audit those.


On Tue, Jun 11, 2013 at 5:24 PM, Eli Friedman <eli.friedman at gmail.com>wrote:

>
>   At a higher level, the big, potentially breaking change here is that
> you're making FunctionProtoType::getArgType() potentially return an
> undecayed type.  There are around 100 users of this API in-tree, and it
> looks like a substantial fraction of them are going to need a similar
> change.  I'm not comfortable with an API named getArgType() returning an
> un-decayed type; even if you catch all the existing cases which need
> additional checks, people writing new code aren't going to be thinking
> about this, and the un-decayed type is almost always the wrong answer from
> a semantic point of view.  Any APIs on FunctionProtoType returning
> un-decayed types should have "un-decayed" in the name, IMO.
>
> http://llvm-reviews.chandlerc.com/D957
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130611/362d289c/attachment.html>