Re: [Emacs-diffs] emacs-25 33ec2ff: Add one more mod-test test

Re: [Emacs-diffs] emacs-25 33ec2ff: Add one more mod-test test

[Ken Brown]

On 11/24/2015 2:14 PM, Eli Zaretskii wrote:
> +  (let ((descr (should-error (mod-test-sum 1 2 3))))
> +    (should (eq (car descr) 'wrong-number-of-arguments))
> +    (should (stringp (nth 1 descr)))
> +    (should (eq 0
> +                (string-match
> +                 (if (eq system-type 'windows-nt)
> +                     "#<module function at \\(0x\\)?[0-9a-fA-F]+ from .*>"
> +                   "#<module function Fmod_test_sum from .*>")
> +                 (nth 1 descr))))

This fails on Cygwin because

   (should-error (mod-test-sum 1 2 3))

evaluates to

   (wrong-number-of-arguments "#<module function at 0x440bd10dc>" 3)

This is like the expected value on MS-Windows, but without the "from". 
Does the missing "from" indicate a bug in the Cygwin build?

Ken

Re: [Emacs-diffs] emacs-25 33ec2ff: Add one more mod-test test

[Eli Zaretskii]

> From: Ken Brown <kbrown@cornell.edu>
> Date: Sat, 28 Nov 2015 21:12:03 -0500
> 
> On 11/24/2015 2:14 PM, Eli Zaretskii wrote:
> > +  (let ((descr (should-error (mod-test-sum 1 2 3))))
> > +    (should (eq (car descr) 'wrong-number-of-arguments))
> > +    (should (stringp (nth 1 descr)))
> > +    (should (eq 0
> > +                (string-match
> > +                 (if (eq system-type 'windows-nt)
> > +                     "#<module function at \\(0x\\)?[0-9a-fA-F]+ from .*>"
> > +                   "#<module function Fmod_test_sum from .*>")
> > +                 (nth 1 descr))))
> 
> This fails on Cygwin because
> 
>    (should-error (mod-test-sum 1 2 3))
> 
> evaluates to
> 
>    (wrong-number-of-arguments "#<module function at 0x440bd10dc>" 3)
> 
> This is like the expected value on MS-Windows, but without the "from". 
> Does the missing "from" indicate a bug in the Cygwin build?

It indicates that your dladdr implementation doesn't work (probably
returns zero).

Please ask on the Cygwin list, and if dladdr is indeed missing, the
test should be updated to accommodate for this return value on Cygwin.

Re: [Emacs-diffs] emacs-25 33ec2ff: Add one more mod-test test

[Ken Brown]

On 11/28/2015 10:42 PM, Eli Zaretskii wrote:
>> From: Ken Brown <kbrown@cornell.edu>
>> Date: Sat, 28 Nov 2015 21:12:03 -0500
>>
>> On 11/24/2015 2:14 PM, Eli Zaretskii wrote:
>>> +  (let ((descr (should-error (mod-test-sum 1 2 3))))
>>> +    (should (eq (car descr) 'wrong-number-of-arguments))
>>> +    (should (stringp (nth 1 descr)))
>>> +    (should (eq 0
>>> +                (string-match
>>> +                 (if (eq system-type 'windows-nt)
>>> +                     "#<module function at \\(0x\\)?[0-9a-fA-F]+ from .*>"
>>> +                   "#<module function Fmod_test_sum from .*>")
>>> +                 (nth 1 descr))))
>>
>> This fails on Cygwin because
>>
>>     (should-error (mod-test-sum 1 2 3))
>>
>> evaluates to
>>
>>     (wrong-number-of-arguments "#<module function at 0x440bd10dc>" 3)
>>
>> This is like the expected value on MS-Windows, but without the "from".
>> Does the missing "from" indicate a bug in the Cygwin build?
>
> It indicates that your dladdr implementation doesn't work (probably
> returns zero).
>
> Please ask on the Cygwin list, and if dladdr is indeed missing, the
> test should be updated to accommodate for this return value on Cygwin.

Thanks.  In fact, dladdr isn't implemented on Cygwin, so dynlib_addr 
always returns false.  I'll update the test.  Should the update be 
Cygwin-specific?  According to http://linux.die.net/man/3/dladdr, dladdr 
is a glibc extension to POSIX, so there could conceivably be other 
systems with the same issue.

Ken

Re: [Emacs-diffs] emacs-25 33ec2ff: Add one more mod-test test

[Eli Zaretskii]

> Cc: emacs-devel@gnu.org
> From: Ken Brown <kbrown@cornell.edu>
> Date: Sun, 29 Nov 2015 10:21:39 -0500
> 
> >> Does the missing "from" indicate a bug in the Cygwin build?
> >
> > It indicates that your dladdr implementation doesn't work (probably
> > returns zero).
> >
> > Please ask on the Cygwin list, and if dladdr is indeed missing, the
> > test should be updated to accommodate for this return value on Cygwin.
> 
> Thanks.  In fact, dladdr isn't implemented on Cygwin, so dynlib_addr 
> always returns false.  I'll update the test.

Thanks.

> Should the update be Cygwin-specific?  According to
> http://linux.die.net/man/3/dladdr, dladdr is a glibc extension to
> POSIX, so there could conceivably be other systems with the same
> issue.

Could be, but I don't know which systems might have this problem.  If
you or someone else have access to *BSD or Solaris, please try that
there.  I wouldn't recommend blindly do that everywhere except
GNU/Linux and Windows, because then we could miss some systems where
the functionality is available, but something needs to be done to have
it.  So I think it's best to fail (pun intended) on the safe side.

Re: [Emacs-diffs] emacs-25 33ec2ff: Add one more mod-test test

[Ken Brown]

On 11/29/2015 11:15 AM, Eli Zaretskii wrote:
>> Cc: emacs-devel@gnu.org
>> From: Ken Brown <kbrown@cornell.edu>
>> Date: Sun, 29 Nov 2015 10:21:39 -0500
>> Thanks.  In fact, dladdr isn't implemented on Cygwin, so dynlib_addr
>> always returns false.  I'll update the test.
>
> Thanks.
>
>> Should the update be Cygwin-specific?  According to
>> http://linux.die.net/man/3/dladdr, dladdr is a glibc extension to
>> POSIX, so there could conceivably be other systems with the same
>> issue.
>
> Could be, but I don't know which systems might have this problem.  If
> you or someone else have access to *BSD or Solaris, please try that
> there.  I wouldn't recommend blindly do that everywhere except
> GNU/Linux and Windows, because then we could miss some systems where
> the functionality is available, but something needs to be done to have
> it.  So I think it's best to fail (pun intended) on the safe side.

OK, I've done this.  I don't have access to *BSD or Solaris, so someone 
else will have to make sure that the test passes on those systems.

Ken