Corinna Vinschen wrote:
> On Mar  3 10:04, Pedro Alves wrote:
>> On Thursday 03 March 2011 09:25:32, Joel Brobecker wrote:
>>>> I'm just wondering, can we assume that the stat structure has always an
>>>> st_ino member?  THere are checks for st_blocks and st_blksize in place
>>>> already.  Is the same required for st_ino?
>>> I think it's fine, at least for now.  The good news is that this
>>> field is described in the Open Group's page for this type:
>>> http://pubs.opengroup.org/onlinepubs/009695399/basedefs/sys/stat.h.html
>>> So I think we can count on it (although, I don't know enough about C
>>> to know how much authority this has)
>>>
>>> But, regardless, I think we can leave it be until such day where
>>> we actually come across a system where the field actually does not
>>> exist... When that happens, we can easily fix the build failure.
>> Can't we just memset `st' instead?
> 
> That would be the simplest solution.  Afterwards, just set the
> non-zero fields.

OK, how about this?