SLIME and CTRL-D

SLIME and CTRL-D

[Peter Keller]

Hello,

I use slime with sbcl on linux and in general it works fine.

My question is: in the raw sbcl repl (say starting sbcl from the command
line) I can do ^D to signal the END-OF-FILE condition for something
reading *standard-input*. SLIME doesn't seem to understand this. How
can I do the equivalent of ^D when using the SLIME repl for sbcl?

Thank you.

-pete

Re: SLIME and CTRL-D

[Helmut Eller]

* Peter Keller [2010-04-08 20:22+0200] writes:

> Hello,
>
> I use slime with sbcl on linux and in general it works fine.
>
> My question is: in the raw sbcl repl (say starting sbcl from the
> command line) I can do ^D to signal the END-OF-FILE condition for
> something reading *standard-input*.  SLIME doesn't seem to understand
> this. How can I do the equivalent of ^D when using the SLIME repl for
> sbcl?

Probably C-u RET.

Helmut

Re: SLIME and CTRL-D

[Peter Keller]

Helmut Eller <eller.helmut@gmail.com> wrote:
> * Peter Keller [2010-04-08 20:22+0200] writes:
> 
>> Hello,
>>
>> I use slime with sbcl on linux and in general it works fine.
>>
>> My question is: in the raw sbcl repl (say starting sbcl from the
>> command line) I can do ^D to signal the END-OF-FILE condition for
>> something reading *standard-input*.  SLIME doesn't seem to understand
>> this. How can I do the equivalent of ^D when using the SLIME repl for
>> sbcl?
> 
> Probably C-u RET.

in a simple test case in slime's repl:

(handle-case (read) (end-of-file () (format t "got eof~%")))

It works just fine. I have yet to test it in my actual application, but
I don't see any reason why it wouldn't work given the test.

Another question is, why C-u RET? I never would have though that did what
it did here. Why does it work?

-pete

Re: SLIME and CTRL-D

[Helmut Eller]

* Peter Keller [2010-04-08 22:29+0200] writes:

> Another question is, why C-u RET? I never would have though that did what
> it did here. Why does it work?

What does C-d in a tty (in cooked mode)?  It sends the possibly empty
input buffer which may be empty.

In Slime, C-d is already taken and there needs to be a way to send the
current input without automatically adding newlines or closing parens,
so C-u RET seemed fitting.

Helmut

Re: SLIME and CTRL-D

[Peter Keller]

Helmut Eller <eller.helmut@gmail.com> wrote:
> In Slime, C-d is already taken and there needs to be a way to send the
> current input without automatically adding newlines or closing parens,
> so C-u RET seemed fitting.

Huh. Thanks for the explanation. I just know C-u to be a "do this N times" 
command, so relating it to stdin was surprising. Then again, I haven't learned
all of the slime commands yet. Maybe I'll run into it in the manual.

Thank you.

-pete

Re: SLIME and CTRL-D

[Jeff Clough]

Peter Keller <psilord@merlin.cs.wisc.edu> writes:

> Helmut Eller <eller.helmut@gmail.com> wrote:
> ... I just know C-u to be a "do this N times" 
> command, so relating it to stdin was surprising.

In Emacs, C-u specifies the prefix argument, which *may* be a repetition
count, but sometimes simply changes the behavior of the command.

I think the canonical example of this is fill-paragraph.  If you give it
a prefix argument, it ignores the value but interprets the presence of
the argument to mean "justify the paragraph too".

Jeff