zwang <nehzgnaw gmail.com> writes:

<code>
class A:B{}
class B:A{}
</code>

and

<code>
interface A:B{}
interface B:A{}
</code>

are silently accepted by dmd.

Thomas Kuehne <thomas-dloop kuehne.thisisspam.cn> writes:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

zwang schrieb am Wed, 20 Apr 2005 16:05:47 +0800:
<code>
 class A:B{}
 class B:A{}
</code>

 and

<code>
 interface A:B{}
 interface B:A{}
</code>

 are silently accepted by dmd.

segfaults under Linux

Added to DStress as
http://dstress.kuehne.cn/nocompile/class_13.d
http://dstress.kuehne.cn/nocompile/class_14.d
http://dstress.kuehne.cn/nocompile/interface_15.d
http://dstress.kuehne.cn/nocompile/interface_16.d

Thomas

-----BEGIN PGP SIGNATURE-----

iD8DBQFCZk3d3w+/yD4P9tIRAihPAJ91qQZ1mwplvzoTKPamHvkjYso/qQCgq/tf
uCGuxxlI24mJZDIfWGgkTCs=
=0glA
-----END PGP SIGNATURE-----

xs0 <xs0 xs0.com> writes:

Thomas Kuehne wrote:
 -----BEGIN PGP SIGNED MESSAGE-----
 Hash: SHA1
 
 zwang schrieb am Wed, 20 Apr 2005 16:05:47 +0800:
 
<code>
class A:B{}
class B:A{}
</code>

and

<code>
interface A:B{}
interface B:A{}
</code>

are silently accepted by dmd.

 
 
 segfaults under Linux
 
 Added to DStress as
 http://dstress.kuehne.cn/nocompile/class_13.d
 http://dstress.kuehne.cn/nocompile/class_14.d
 http://dstress.kuehne.cn/nocompile/interface_15.d
 http://dstress.kuehne.cn/nocompile/interface_16.d
 
 Thomas

Hmm, if you think about it, it could actually be allowed (interfaces 
only, classes have an overload problem) - if A is a B and B is an A, 
obviously A equals B, and they could be merged, while both A and B 
become aliases to the merged interface.

OTOH, I see no absolutely no use this could have, and it's far more 
likely that a bug in the code is involved..


xs0