[Lsb-infrastructure] devchk

Denis Silakov silakov at ispras.ru
Fri Jul 6 07:50:43 PDT 2007


Wichmann, Mats D wrote:
> Denis Silakov wrote:
>   =

>> Wichmann, Mats D wrote:
>>     =

>>> so, at least as things stand at the moment, a generic type in
>>> addition to arch-specific types seems "more correct", at least
>>> if that type is used in function prototypes, right?
>>>
>>>       =

>> Yes, that's true.
>>     =

>
>
> So looking a little further afield, it looks like devchk
> is no longer getting the structure member size/offset
> checks generated and the structure size check (except for
> a generic version).   I'm not remembering if this was
> intentional or if it's something that broke in the
> transition away from ArchTypeMem....
>
> I've looked at sys_stat_h.c, sys_statfs_h.c and
> sys_statvfs_h.c; all three have some emtpy architecture
> #if defined sections, but no code for the corresponding
> structures excepting that there's arch-generic code
> to check the size of statfs. Of the six structures
> (stat, statfs, statvfs and corresponding *64), only
> the statfs* ones have a generic entry in addition to
> the arch entries, which matches the discussion so
> far on the topic. =

>
>   =

This situation came from patch from bug 1507
(http://bugs.linuxbase.org/show_bug.cgi?id=3D1507)m which is partially
merged at the moment. The patch suggested not to test types whose base
types are not included in the LSB, but it didn't take into account that
some types don't have base types at all (formally they have base type
with zero id, and there is no type with such id in the database).

A new patch is already available within the bug 1507.

-- =

Regards,
Denis.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.linux-foundation.org/pipermail/lsb-infrastructure/attachm=
ents/20070706/b3948136/attachment.htm


More information about the Lsb-infrastructure mailing list