check-expect snip output unfriendly

[jasonhemann]

The following error-scenario seems unnecessarily unfriendly, and I'm wondering if something nicer is possible.

A student might write the following test program:

#lang htdp/bsl
(require 2htdp/image)
(check-expect (circle 10 'outline 'black) 'o)

This test will fail. The student receives an error message containing a snip that contains the expected value. Our student then---mistakenly, accidentally, or what have you---then picks up the whole snip and drops that in as a replacement expect value.

Now that student sees the even more confusing:

Actual value . differs from (instantiate (class ...) ...), the expected value.

This seems especially injurious because it shows up when a student is already grappling with code that isn't behaving the way he expects. I for one do not know how to explain this behavior in a way that makes good sense to a BSL student. Is something nicer is possible?

I do not know if this properly belongs here, or in the htdp repo, or elsewhere.

[rfindler]

I'm not sure what's the best fix for this either, but one possibility might be an error saying that the box around the image snip is not a legal program in BSL. If that sounds right, then I think that that'd be a change in the bsl repo somewhere.

[mfelleisen]

I get

Welcome to DrRacket, version 8.13.0.9 [cs]. Language: htdp/bsl, with debugging [custom]. The test passed!

for ``` #lang htdp/bsl (require 2htdp/image) (check-expect (circle 10 'outline 'black) .) ``` when I copy over the black circle. Why would a student copy-and-paste the square when he expects to see a black circle?

[sorawee]

Why would a student copy-and-paste the square when he expects to see a black circle?

Perhaps the student might not know that it is possible to copy only just the circle, and thought that the frame is just for decoration. I could see myself easily make this mistake.

[mfelleisen]

So `(if (image? x) x (frame (render x)))` ?

[rfindler]

FWIW, I agree with Sorawee that what Jason describes is a totally reasonable (although wrong) thing to do and it would be good if something better happened in response to it. That said, I don't get what Matthias is trying to say with the if expression.

[jasonhemann]

I don't quite understand the suggested fix yet either.

[sorawee]

My guess is that Matthias was suggesting a workaround to make the framing snip comparable as an image?

[mfelleisen]

That's exactly what my if expression was supposed to suggest. [[ Sorry I was on the road for the week. ]]

[shhyou]

If I got it right, it looks like the framed thingy is an editor% snip here.
https://github.com/racket/htdp/blob/8037f1cc620ed91fe68362ce5ad3d887269eb725/htdp-lib/test-engine/markup-gui.rkt#L145-L152
Is it possible for the copy behavior of the snip be changed to copying its content instead? (Or perhaps changing the pasting behavior. IIRC some snips do this.)

[rfindler]

Oh that's a great idea! (We would make a subclass and do that only for testing ones, I suppose.) Robby

…

On Wed, Jul 3, 2024 at 4:15 PM shuhung ***@***.***> wrote: If I got it right, it looks like the framed thingy is an editor% snip here. https://github.com/racket/htdp/blob/8037f1cc620ed91fe68362ce5ad3d887269eb725/htdp-lib/test-engine/markup-gui.rkt#L145-L152 Is it possible for the copy behavior of the snip be changed to copying its content instead? — Reply to this email directly, view it on GitHub <#673 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AADBNMGGL52OQHYMJLHABJ3ZKRSYRAVCNFSM6AAAAABJ2M6PMWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMBXGMYTIMZTGQ> . You are receiving this because you commented.Message ID: ***@***.***>