>>>>> "Balint" == Balint Reczey <balint.rec...@canonical.com> writes:
Balint> Hi Sam, Thank you for the quick response. >> That said, I think libverto in Debian should support all the >> options, and certainly should support libevent. That won't make >> things easier for Ubuntu really, if they want to avoid building >> libverto against libev, but it will let Debian users use libevent >> if that's what they want. Balint> In general I agree with offering the choice, but in that Balint> particular package's case I saw no prior indication of any Balint> user wanting to use libverto-libevent1 rather than Balint> libverto-libev1. I see that you maintain krb5, too. What Balint> advantage do you see in providing libverto-libevent1, too, Balint> that would make the extra complexity worth it? So, if you are using libverto, in a particular application, you are probably happier if your libverto event loop matches up well with your application event loop. So if you're using a library like krb5 in an application that uses libevent or glib, you probably would like libverto to use that event loop. (libverto is effectively designed to allow libraries to use an event loop.) This doesn't matter in practice because not a lot of people use libverto. And because libkrb5 doesn't tend to use libverto much; I seem to recall it's more a kdc-side thing. In general in Debian though we tend to turn on all the upstream compile-time options that a package supports. Honestly, libverto seems like one too many layers for my taste. Yeah, it's kind of a PITA to glue too event loops together, but it actually is possible in the cases that matter, and I kind of wish that krb5 had just chosen one event loop (pick any one) rather than going and designing an abstraction for abstracting away the choice of event loop. So, with my Debian-packages-should-support-upstream options hat on, I kind of feel like libverto should support all the options as a wishlist priority. But I can't say that I think it would make much difference or that it's where I plan to spend a lot of time.