hg: lambda/lambda/langtools: Next round of implementation reflecting the latest 'State of the Lambda' draft; implemented features are:

[Nathan Bryant]

Right, because (just as scala.Nothing is a subtype of every type)
lambdas are widening-convertible to every compatible SAM type. I get
that.

The point was that the call site:

(1) foo({x -> something})

is inherently different from the call site:

(2) foo(SomeTaggingInterface {x -> something})

If we infer that (1) is the same as (2) in the presence of an overloaded
foo(), we've added a new source of surprise to the language. Maybe it's
not the most common use case, but there is no such opportunity for
surprise with tagging interfaces today.

-----Original Message-----
From: neal.gafter at gmail.com [mailto:neal.gafter at gmail.com] On Behalf Of
Neal Gafter
Sent: Monday, July 26, 2010 4:48 PM
To: Nathan Bryant
Cc: Neal Gafter; Joe Darcy; maurizio.cimadamore at oracle.com;
lambda-dev at openjdk.java.net
Subject: Re: hg: lambda/lambda/langtools: Next round of implementation
reflecting the latest 'State of the Lambda' draft; implemented features
are:

The "more specific" analysis doesn't depend on knowing the argument
types.

On Monday, July 26, 2010, Nathan Bryant <nathan.bryant at linkshare.com>
wrote:
> The usual rules would seem to be appropriate when the parameter type
is known. In this case, it is not.