[ofa-general] [PATCH 1/3 v2] libibmad. 2nd revision for WinOF portablity.
Sasha Khapyorsky
sashak at voltaire.com
Wed Jan 21 10:36:42 PST 2009
On 07:51 Wed 21 Jan , Sean Hefty wrote:
>
> I think all of this is overkill. Is it really that difficult to ensure that new
> fields are placed into the array correctly, especially compared to the cost,
> say, of ensuring that the field offsets are correct or maintaining an additional
> pre-processing script?
For maintenance troubles start I will need to ensure that all fields in
discussed array are placed in proper order even before the patch - there
was never such requirement before and there are more than 200 fields. :(
But my concern is even more general - if not using c99 initializations
for arrays and structures will become a general requirement I will not
happy - what is the point to have a good development tools with all those
nice and useful things if we cannot use them? And finally I'm not about
exotic extensions, it is standard stuff...
BTW, did somebody try to raise this issue in WinOF? What is the reason
to be closed with VC?
> We're reducing the broader maintenance costs overall by
> having a single code base.
>
> Is this is only open issue integrating the ib-mgmt WinOF support changes?
Yes, as far I as can see after brief review. Likely I will apply 1 and 3
later today.
Sasha
More information about the general
mailing list