[PATCH] Fixing source range for functional casts and unresolved ctor calls.

[Enea Zaffanella]

On 07/05/2013 08:54 PM, Eli Friedman wrote:
> On Fri, Jul 5, 2013 at 12:45 AM, Enea Zaffanella <zaffanella at cs.unipr.it
> <mailto:zaffanella at cs.unipr.it>> wrote:
>
>     Please find attached a patch (including testcase) for fixing the
>     computation of source ranges for a couple of AST nodes.
[...]
> Do you actually need a definition of initializer_list for this test?
>   IIRC, you can get an CXXUnresolvedConstructExpr with any dependent
> type.  I like using unit-tests to check source ranges rather than
> checking the output of -ast-dump.
>
> Otherwise, LGTM.
>
> -Eli

Committed in 185773 with simplified testcase and no (ab-)use of 
-ast-dump, as suggested.

Enea.