On Tue, Mar 23, 2010 at 10:34 AM, Dave Shield
<[email protected]>wrote:
> On 22 March 2010 18:46, Bart Van Assche <[email protected]> wrote:
> > How well have these patches been tested, and on which platforms ?
>
> The first three were in response to queries on the mailing list.
> I've tested (2) and (3) myself (on Fedora), and received
> positive feedback from the original posters.
> (4) has also been tested on my Fedora setup
> (1) is based on a back-port of fixes from the 5.5.x branch
> Unfortunately I don't have multi-Terabyte disks myself
>
>
> > * (1) will only work when sizeof(long) > sizeof(int).
>
> But what would be the effect if long (and int) are both 32-bit?
> My understanding is that the comparison (long_ret > MAX_INT_32)
> would never match, so the latched value wouldn't be assigned.
> Hence the result returned would be the same as before.
>
Personally I consider the fact that the patch does not have the same effect
on 32-bit systems as on 64-bit systems as a flaw.
There is yet another shortcoming in the patch that I hadn't mentioned yet:
limiting values to 0x7fffffff introduces a regression for those platforms
where sizeof(int) == 8. While reporting large (>= 2**31) disk statistics
currently works fine for those platforms, this patch breaks it.
> > * (4) breaks the Windows build because netsnmp_get_mib_directory() is
> not
> > exported from the Net-SNMP DLL on the 5.4 branch (see also
> >
> http://net-snmp.svn.sourceforge.net/viewvc/net-snmp?view=rev&revision=18347
> ).
>
> OK.
> What if:
> a) 'netsnmp_get_mib_directory' was added to
> win32/libsnmp_dll/libsnmp.def ?
> or
> b) The update to snmp/agent.c was omitted?
> (i.e. just issue the message when "module not found"
> errors are printed)
>
Both approaches seem OK to me. Can you please post an updated patch ?
Bart.
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
