In article <FA654396-DF20-11D9-90E9-000502EE4B47@libero.it>,
Riccardo Mottola  <zuse@libero.it> wrote:

I have a centris lying somewhere and I will try it myself...

christos

>Hey,
>
>since I failed to get a usable box ith the netbsd 2.0.2 release I 
>thought that trying the current netbsd-3 snapshot could do no damage.
>
>But it is really in bad, bad shape.
>
>My box is a a MacIIci, 72MBytes of ram and a 2Gb disk. I installed 
>directly there instead than on my test disk...
>
>I got the snapshot from netbsd-daily, netbsd-3
>
>1. the sysint procedure is still broken with regard to newfs and mount 
>to /targetroot. I used the usual trick
>
>2. the kernel seems to have problems, if the sysinst wasn't successful 
>or got restarted for some reason, it would kill itself giving usually no 
>space left on disk message and a reboot from shell causes a kernel 
>panic. Sysinst has to go in one single step, restarting from the 
>beginning for any reason will cause deep trouble. I assume kernel 
>problems.
>
>3. most of the packages didn't even uncompress properly! sometimes pax 
>returns a "killed by signal" but most often the console gets swamped 
>with tar and checksum errors. this with various of the .tgz (man, etc, X 
>packages, comp...). I didn't delete the packages.
>
>I ignored the errors, had it continue so that it built devices and I 
>could reboot, indeed the system reboots! Miracle! But X11 will fail for 
>example due to broken library symbols (I guess indeed that the tars 
>weren't uncompressed properly)
>
>I tried to uncompress comp.tgz again now from the running system and it 
>appeared again that the tar was invalid, after some uncompressing it 
>started to search repeatedly for new headers (typical corrupted tars). 
>Now it is a question if the .tgz are corrupted, the tar program itself 
>or the kernel...
>
>-R
>
>