Fwd: Testing Alpha Vary

[Jay McCarthy]

---------- Forwarded message ----------
From: Jay McCarthy <j...@cs.byu.edu>
Date: Mon, Nov 28, 2011 at 7:10 PM
Subject: Re: Testing Alpha Vary
To: Logan Kearsley <chrono...@gmail.com>


You should describe what the results should be in English, or write something like

(alpha-vary ...) "should be like" (fun (x271) x271)

Jay

On Mon, Nov 28, 2011 at 4:32 PM, Logan Kearsley <chrono...@gmail.com> wrote:

So, alpha-vary returns an expression with different unique IDs every
time, which means I can't put anything static in the test cases. So,
how should we do tests on alpha-vary?

-l.

--
Jay McCarthy <j...@cs.byu.edu>
Assistant Professor / Brigham Young University
http://faculty.cs.byu.edu/~jay

"The glory of God is Intelligence" - D&C 93

--
Jay McCarthy <j...@cs.byu.edu>
Assistant Professor / Brigham Young University
http://faculty.cs.byu.edu/~jay

"The glory of God is Intelligence" - D&C 93

[Joey Hewitt]

Can we convert the gensym'd symbol(s) to a string and use a regular
expression to make sure it "looks like" a non-alpha-vary'd
expression? A potential problem is that it relies on unspecified
behavior from gensym, but is that something you would dock for?

Joey

On Nov 28, 7:10 pm, Jay McCarthy <j...@cs.byu.edu> wrote:
> ---------- Forwarded message ----------
> From: Jay McCarthy <j...@cs.byu.edu>
> Date: Mon, Nov 28, 2011 at 7:10 PM
> Subject: Re: Testing Alpha Vary
> To: Logan Kearsley <chronosur...@gmail.com>
>
> You should describe what the results should be in English, or write
> something like
>
> (alpha-vary ...) "should be like" (fun (x271) x271)
>
> Jay
>

> On Mon, Nov 28, 2011 at 4:32 PM, Logan Kearsley <chronosur...@gmail.com>wrote:
>
> > So, alpha-vary returns an expression with different unique IDs every
> > time, which means I can't put anything static in the test cases. So,
> > how should we do tests on alpha-vary?
>
> > -l.
>
> --
> Jay McCarthy <j...@cs.byu.edu>

> Assistant Professor / Brigham Young Universityhttp://faculty.cs.byu.edu/~jay

[Jay McCarthy]

That's not possible to do.

You could write something like

equal-upto-alpha?

like what I've given you for types, but it will be more complicated

[Joey Hewitt]

Oh, because it's hard to make sure the same symbol used in multiple
places was varied the same way in all instances. Right?

Joey

[Jay McCarthy]

Yup

[Wayne Robison]

Should our tests for generate-constraints be similar, since the label ids are going to be unique each time?

--Wayne

[Jay McCarthy]

Ya, although you can use my constraint-list=? function

On Thu, Dec 1, 2011 at 12:31 PM, Wayne Robison <reallybla...@gmail.com> wrote:

Should our tests for generate-constraints be similar, since the label ids are going to be unique each time?

--Wayne