On Fri, Feb 24, 2012 at 10:38 AM, Michal Privoznik <mprivozn(a)redhat.com> wrote:
On 24.02.2012 10:19, Stefan Hajnoczi wrote:
Thank you for the opportunity. I was personally thinking about something
libvirt-snmp related. Nowdays, it is difficult to add new elements to
MIB, as some parts of code were generated by mib2c. Any change to MIB
requires regeneration of such source files and thus leads to loss of all
previous changes. So one thing that is coming to my mind is drop this
dependency and use libsnmp directly. But I am not sure it is worth of GSoC.
If this project is self-contained and can be completed in 12 weeks by
a person without prior libvirt and SNMP experience, but fluent C
programming skills, then it sounds suitable.
Is there any creativity required or problems to solve that aren't
grunt work? For example, if you just need to run mib2c and then
manually diff to produce the final C version, then this sounds like a
lot of manual work but little gain for the student.
Stefan