On Sun, Apr 24, 2016 at 6:34 AM, Daniel Eischen <deischen_at_freebsd.org> wrote: > On Sat, 23 Apr 2016, Warner Losh wrote: > > On Sat, Apr 23, 2016 at 7:51 AM, Daniel Eischen <deischen_at_freebsd.org> >> wrote: >> >> [CC trimmed] >>> >>> On Fri, 22 Apr 2016, Warner Losh wrote: >>> >>> >>>> I personally will be refraining from engaging further. I plan on seeing >>>> what gaps there are by adding support to NanoBSD for packages. I'll be >>>> busy >>>> with that. In talking to Glen and others, we've already identified a few >>>> easy gaps to fill. Once they've done that, I'll get going on NanoBSD >>>> with >>>> the goal to be able to use it to build a bootable system of any >>>> architecture from packages with no root privs. I expect to find issues, >>>> but >>>> I don't expect to find any issue that's intractable. I expect after the >>>> issues are resolved, the end product will be better for everybody. >>>> >>>> >>> Thank you for working on NanoBSD. Do you think it would be possible >>> to add support for optionally building dump(8) images instead of dd? >>> >> >> >> What do you mean by that, exactly? It would be relatively easy to add >> a step that runs dump on the _.disk.image file and squirrel that away. >> Last orders the code currently calls it, I believe. Is it something as >> simple >> as this, or is there some more complexity that I'm failing to understand >> or grasp? >> > > Perhaps I'm missing something, but when last_orders() is called, > isn't the disk already unmounted and 'mdconfig -d -u' already > run? dump 0f - _.disk.image > ~/foo.dump worked for me just now. Is there some reason that it wouldn't work for you in last orders if you tossed a NANO_DISKIMGDIR in front of it and last orders would work for you. You could even pipe it into some compression program... WarnerReceived on Sun Apr 24 2016 - 14:16:38 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:04 UTC