any chance we will get tail calls in C# 4.0?

[not_a_commie]

A tail call is a way of removing the current method from the stack
before recursing. See http://blogs.msdn.com/shrib/archive/...25/360370.aspx
.. I would love to see it in C# 4.0 and wouldn't mind if we needed a
special keyword to make it happen.

[Peter Duniho]

Re: any chance we will get tail calls in C# 4.0?

On Thu, 06 Nov 2008 10:30:52 -0800, not_a_commie <notacommie@gma il.com>
wrote:

A tail call is a way of removing the current method from the stack
before recursing. See
http://blogs.msdn.com/shrib/archive/...25/360370.aspx
. I would love to see it in C# 4.0 and wouldn't mind if we needed a
special keyword to make it happen.

I doubt you'd see that. Tail recursion is so trivially implemented as a
plain iterative loop, I can't imagine that such a feature would meet the
bar for inclusion in the language.

I see why MSIL might have "tail" for supporting functional languages, but
this doesn't seem like something that needs to be exposed in C#. Any code
that could take advantage of it really just should be written differently
anyway.

Pete

[Ben Voigt [C++ MVP]]

Re: any chance we will get tail calls in C# 4.0?

Peter Duniho wrote:

On Thu, 06 Nov 2008 10:30:52 -0800, not_a_commie
<notacommie@gma il.comwrote:
>

>A tail call is a way of removing the current method from the stack
>before recursing. See
>http://blogs.msdn.com/shrib/archive/...25/360370.aspx
>. I would love to see it in C# 4.0 and wouldn't mind if we needed a
>special keyword to make it happen.

>
I doubt you'd see that. Tail recursion is so trivially implemented
as a plain iterative loop, I can't imagine that such a feature would
meet the bar for inclusion in the language.

I think more generally, switch/goto case (there's no requirement that the
tail call has to be recursion back to the same function, all participating
functions would have to be combined and turned into different case blocks in
a single master function).

>
I see why MSIL might have "tail" for supporting functional languages,
but this doesn't seem like something that needs to be exposed in C#. Any
code that could take advantage of it really just should be
written differently anyway.
>
Pete

[Michael D. Ober]

Re: any chance we will get tail calls in C# 4.0?

"not_a_comm ie" <notacommie@gma il.comwrote in message
news:216e1a14-8e20-4ab3-bf39-4d58c4d4704d@r3 7g2000prr.googl egroups.com...

>A tail call is a way of removing the current method from the stack
before recursing. See
http://blogs.msdn.com/shrib/archive/...25/360370.aspx
. I would love to see it in C# 4.0 and wouldn't mind if we needed a
special keyword to make it happen.
>

This should be a compiler optimization as it's very easy to detect and even
easier to implement once detected.

Mike.