Just a passing thought - Test::Import::* suggests the import of data,
rather than code.
The generic name for importing code into the current namespace would
be 'include' I guess, so how about
Test::Include::DontRun
?
On 7 October 2010 10:11, Charles Colbourn
<charles.colbo...@googlemail.com> wrote:
> ---------- Forwarded message ----------
> From: Charles Colbourn <charles.colbo...@googlemail.com>
> Date: 7 October 2010 10:10
> Subject: Re: Module naming - Test::Import::DontRun
> To: ebhans...@cpan.org
>
>
> @Eirik
>
>>>
>>> There's some complex and potentially slightly
>>> fragile stuff to handle @EXPORT if you want to, but the basic case
>>> works pretty well.
>>
>> Whoops. What would your legacy script(?) use @EXPORT for?
>
> Some of the code I'm testing in our own legacy codebase includes
> modules that 'do things' when use'd or require'd, and also export
> methods via @EXPORT. My code attempts to handle this by duplicating
> the @EXPORT* declarations outside the anonymous scope. It does this
> simply by parsing the code and finding the appropriate strings. You
> can switch this behaviour off if need be, although I'm tempted to make
> it default to 'off' since it may not be as robust as the rest.
>
>
>>
>> (I'd be more concerned about __END__, __DATA__, __FILE__, and __LINE__.
>> Okay, the latter two could be handled with a #line directive, but to handle
>> the first two right, it seems you'd have to parse Perl ... or drop the {}s?)
>>
> Eirik++ I'd overlooked those. I'll see what I can do with it, write
> some test cases etc.
>
>
> @Nadim
>
> "OK, I get what you want to do. I have seen this before. My brains being what
> they are lately, I don't remember where but it was not so long ago. Maybe Andy
> maybe someone else. Maybe even you ;)
>
> So before you go further, you can, if you so wish, look around a bit."
>
> I've had a dig around CPAN, and sifted through 'Perl testing, a
> developers notebook' and not found anything. I think I did bandy this
> idea around about 6 months ago, but the version I had then depended on
> modifying the legacy code, and was a glorified version of:
> {
> if (caller()){last}
> #code
> }
>
> So perhaps that was it? I don't think I posted here about it, but I
> may have mentioned it in the CB on Perlmonks, and I definitely floated
> the idea past a couple of people. OTOH if there's something around
> that does this already, that would save me some work :-)
>
