Johan Ihrén <johani%johani.org@localhost> writes: > And here: > >> I would say that if we can get things to build w/o pbulk that would be >> the most important step, and debugging the not-pbulk/pbulk secondary. >> Right now it's hard to recommend 4.1. > > Unless I'm missing something it sounds like you just answered your own > question above. [about 3.3] 4.2 builds. And 4.1 built pretty recently - I built it in January without trouble (because I have a binary package from them and don't remember issues). > So, FWIW, I'm still running 3.3 on a number of machines. While the > problems of building 4.1 tools used to be one reason the other catch > was a large chunk of custom software that I just recently finished > porting to 4.X. I still intend to retain 3.3 for a year before fully > trusting 4.X to be stable enough. Yes, I am that cautious. >> any hope of PCI passthrough on 4.1/4.2? >>> >>> Someone needs to convert the python patch we had for 31 for 4.1. >>> Not sure what's needed for 4.2. >> >> That would be really great to have, because then we could recommend 4.1 >> period and 3.1/3.3 could become really deprecated. > > Please don't be too quick in deprecating 3.3. If we could keep Xen2 > around for several years after Xen4 arrived, surely we can keep Xen3 > for while. I didn't mean to propose removing it from pkgsrc. I meant more to recommend "Don't use this unless you have some specific reason". That's been true of xen2 for a really long time, and it hasn't been deleted yet.
Attachment:
pgp2XT66EmgGx.pgp
Description: PGP signature