Need help with catching reader error

Need help with catching reader error

[Mike Gran]

Hi-

Here's a puzzle for you.

I want to write a test for the test suite to catch a lexical syntax
error, like the following non-existent named character.  But I need to
somehow introduce another layer of evaluation.  In the following, Guile
would tell me that my test script has an error and then quit, instead of
catching the error and moving on.

(use-modules (test-suite lib))

(define exception:read-error
  (cons 'read-error "^.*"))

(with-test-prefix "basic char handling"
 (pass-if-exception "non-existent named character" 
		    exception:read-error
		    #\foobar))

Thanks,

Mike Gran

Re: Need help with catching reader error

[Andreas Rottmann]

Mike Gran <spk121@yahoo.com> writes:

> Hi-
>
> Here's a puzzle for you.
>
> I want to write a test for the test suite to catch a lexical syntax
> error, like the following non-existent named character.  But I need to
> somehow introduce another layer of evaluation.  In the following, Guile
> would tell me that my test script has an error and then quit, instead of
> catching the error and moving on.
>
> (use-modules (test-suite lib))
>
> (define exception:read-error
>   (cons 'read-error "^.*"))
>
> (with-test-prefix "basic char handling"
>  (pass-if-exception "non-existent named character" 
> 		    exception:read-error
> 		    #\foobar))
>
You have a syntax error in your source script, which will hence not be
able to be read and consequently not be able to be evaluated.

Use this instead of #\foobar in the source text:

(with-input-from-string "#\\foobar" read)

This way, your source has legal syntax.

--Rotty

Re: Need help with catching reader error

[Mike Gran]

On Wed, 2009-04-22 at 16:53 +0200, Andreas Rottmann wrote:
> Mike Gran <spk121@yahoo.com> writes:
> > (with-test-prefix "basic char handling"
> >  (pass-if-exception "non-existent named character" 
> > 		    exception:read-error
> > 		    #\foobar))
> >
> You have a syntax error in your source script, which will hence not be
> able to be read and consequently not be able to be evaluated.
> 
> Use this instead of #\foobar in the source text:
> 
> (with-input-from-string "#\\foobar" read)
> 
> This way, your source has legal syntax.

> --Rotty

Nice.  It works great.  Thanks.

-Mike Gran

Re: Need help with catching reader error

[Ludovic Courtès]

Hi,

Andreas Rottmann <a.rottmann@gmx.at> writes:

> Use this instead of #\foobar in the source text:
>
> (with-input-from-string "#\\foobar" read)
>
> This way, your source has legal syntax.

Yes, there are other examples of this in `readers.test'.

Thanks,
Ludo'.