On Sun, Mar 27, 2011 at 8:28 AM, Magnus Fromreide <[email protected]>wrote:

> On Sat, 2011-03-26 at 19:29 -0400, Bill Fenner wrote:
> > On Mar 26, 2011, at 4:02 PM, Magnus Fromreide <[email protected]>
> wrote:
> >
> > > On Sat, 2011-03-26 at 14:08 -0400, Bill Fenner wrote:
> > >> Hi,
> > >>
> > >> I tried a couple of recent svn checkouts (most recently, r20159), and
> > >> found that net-snmp-features.h isn't installed via "make install":
> > >>
> > >>
> > >> In file included from /Users/fenner/tmp/include/net-snmp/types.h:421,
> > >>                 from
> /Users/fenner/tmp/include/net-snmp/definitions.h:22,
> > >>                 from
> /Users/fenner/tmp/include/net-snmp/net-snmp-includes.h:67,
> > >>                 from subagent.c:7:
> > >> /Users/fenner/tmp/include/net-snmp/library/snmp_api.h:33:40: error:
> > >> net-snmp/net-snmp-features.h: No such file or directory
> > >>
> > >>
> > >> I assume that this is just an oversight, and everyone who's building
> > >> stuff recently is just working inside the source tree.
> > >>
> > >> Manually copying net-snmp-features.h from includes/net-snmp to
> > >> /Users/fenner/tmp/include/net-snmp/ was sufficient.  Sadly, I haven't
> > >> yet figured out where the Makefile is that does the header
> > >> installation or I would supply a diff, but hopefully this is enough
> > >> info for a developer to fix this for others.
> > >
> > > I would suggest that the attached patch is the most correct one but it
> > > is problematic in it's own right.
> > >
> > > The effects are that
> > >
> > > 1) The statistics functions will always exist
> > > 1a) Thus the header won't have to be installed and so Bill's problem is
> > >    solved
> > > 2) The return value of snmp_increment_statistic* is changed to 1 if the
> > >   counter existed, 0 otherwise. The new value of the counter is not
> > >   available any more (Before it was problematic as it returned false
> > >   fails once in a while).
> > >   The return value of snmp_get_statistic will always be 0
> > > 3) The functions will always be provided as stub functions.
> > >
> > > Is this patch acceptable?
> > >
> > > /MF
> >
> > Sorry to be dense - it seems to me that changing the right Makefile to
> > install this header is the easiest and is not obviously wrong. What am
> > I missing?
>
> The ABI effect. Should the symbols be present in the library or not?
>

Shouldn't <net-snmp/net-snmp-features.h> be considered as part of the ABI ?
I think it should be installed too.

Bart.
------------------------------------------------------------------------------
Enable your software for Intel(R) Active Management Technology to meet the
growing manageability and security demands of your customers. Businesses
are taking advantage of Intel(R) vPro (TM) technology - will your software 
be a part of the solution? Download the Intel(R) Manageability Checker 
today! http://p.sf.net/sfu/intel-dev2devmar
_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders

Reply via email to