Re: srfi-64 apathetic test-error

[Taylan Kammer <taylan.kammer@gmail.com>]

Sorry, I said something wrong in my previous email.

The SRFI-64 specification does not mention the possibility of providing
an arbitrary predicate.  That's a feature supported by my R7RS-based
implementation of SRFI-64:

    https://github.com/TaylanUB/scheme-srfis/tree/master/srfi/64

(The file 'execution.body.scm' contains most of the logic.)

I've tried to figure out what the reference implementation supports but
the code is overly complicated.  There's a comment saying "TODO: decide
how to specify expected error types for Guile" which makes me think it
doesn't support Guile that well.  I think it doesn't actually use the
"error-type" information at all in the case of Guile and just matches
all errors it catches and considers them correct.  See lines 800 to 818
here:


http://git.savannah.gnu.org/cgit/guile.git/tree/module/srfi/srfi-64/testing.scm

Last time I suggested Guile to adopt my SRFI-64 implementation, it was
argued against on the grounds that it's not been around as long as the
reference implementation and therefore possibly less stable.  Maybe I
should try again, because IMO the reference implementation is written in
a terrible style that makes it very difficult to comprehend and improve.

- Taylan


On 24.10.2019 18:18, Taylan Kammer wrote:
> On 23.10.2019 06:37, Christopher Howard wrote:
>> Hi, I was just wondering: `test-error' from (srfi srfi-64) gives a pass
>> if it catches an exception, but it doesn't seem to care whether or not
>> the exception caught matches what you told it to expect. Is this a bug?
> 
> The (test-error ...) form takes an optional "error-type" operand.  It
> can be an arbitrary predicate (one-argument procedure that returns a
> Boolean value) which tests if the error is the expected kind.
> 
> See here for a complete documentation of the SRFI-64 API:
> 
>     https://srfi.schemers.org/srfi-64/srfi-64.html
> 
> - Taylan
>