[type_traits] Borland 2006 tests

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

[type_traits] Borland 2006 tests

zara-7
I am working for bcbboost, and I have begun making regression test,
going from basics upwards.

I have found  (with my current configuration) therer *hundreds* of
type_traits library tests that fail. I am trying to find a solution to
those, but I cannot keep from wondering, was BCB6 also failing, and it
is not important? Was it failing and nobody noticed? Am I failing to
see the evident, and all my work process is poisoned form the
beginning?

To give an example, remove_const fails for the following cases:

2: volatile -> volatile
3: const volatile -> volatile
7: *volatile -> *volatile
8: *const volatile -> *volatile
13: volatile[2] -> volatile[2]

As I have seen on other threads, there might be an issue with the
order of quialifications, so maybe the soltunio is to redefine the
test macros, but I hope someone will give me some idea to continue
with minimum effort.

TIA, bes regards,

Zara

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [type_traits] Borland 2006 tests

Nicola Musatti
Zara wrote:
> I am working for bcbboost, and I have begun making regression test,
> going from basics upwards.
>
> I have found  (with my current configuration) therer *hundreds* of
> type_traits library tests that fail. I am trying to find a solution to
> those, but I cannot keep from wondering, was BCB6 also failing, and it
> is not important? Was it failing and nobody noticed? Am I failing to
> see the evident, and all my work process is poisoned form the
> beginning?

If you are talking about the regression tests available in 1.33.1 the
current results for bcbboost CVS are the same for BCB2006 and BCB6,
about 50 test passed and 17 failed. You can browse the results here:
http://bcbboost.sourceforge.net/test/current/cs-win32.html

Cheers,
Nicola Musatti

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [type_traits] Borland 2006 tests

zara-7
On Wed, 01 Mar 2006 22:21:35 +0100, Nicola Musatti
<[hidden email]> wrote:

>Zara wrote:
>> I am working for bcbboost, and I have begun making regression test,
>> going from basics upwards.
>>
>> I have found  (with my current configuration) therer *hundreds* of
>> type_traits library tests that fail. I am trying to find a solution to
>> those, but I cannot keep from wondering, was BCB6 also failing, and it
>> is not important? Was it failing and nobody noticed? Am I failing to
>> see the evident, and all my work process is poisoned form the
>> beginning?
>
>If you are talking about the regression tests available in 1.33.1 the
>current results for bcbboost CVS are the same for BCB2006 and BCB6,
>about 50 test passed and 17 failed. You can browse the results here:
>http://bcbboost.sourceforge.net/test/current/cs-win32.html
>

Thank's for the link, I had not noticed it. BTW, does this mena we
must not try to pass the tests, they are impossible to pass?
Because if Borland is unable to pass the type_traits test (at least
the non-tricky ones), then many other libraries won't work.


Checking the results:

 I find is_union is markes 'Pass'g, but it is simply disconnected with
a warning, it still seems so.

I will test my config, as is_class fails on my tests.

And with my config tricy_is_enum is 'Pass', not 'Fail'

regards,

Zara

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: [type_traits] Borland 2006 tests

zara-7
In reply to this post by Nicola Musatti
On Wed, 01 Mar 2006 22:21:35 +0100, Nicola Musatti
<[hidden email]> wrote:

>Zara wrote:
>> I am working for bcbboost, and I have begun making regression test,
>> going from basics upwards.
>>
>> I have found  (with my current configuration) therer *hundreds* of
>> type_traits library tests that fail. I am trying to find a solution to
>> those, but I cannot keep from wondering, was BCB6 also failing, and it
>> is not important? Was it failing and nobody noticed? Am I failing to
>> see the evident, and all my work process is poisoned form the
>> beginning?
>
>If you are talking about the regression tests available in 1.33.1 the
>current results for bcbboost CVS are the same for BCB2006 and BCB6,
>about 50 test passed and 17 failed. You can browse the results here:
>http://bcbboost.sourceforge.net/test/current/cs-win32.html
>
>Cheers,
>Nicola Musatti


<Addendum to my last message>

tricky_is_enum I do not know if it is 'Pass' or 'Fail', I have
detected a bug in my regression test fixture, and it was not tested.

Zara

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost