On 5/31/16 4:51 PM, Bryan Drewery wrote: > On 5/31/16 4:41 PM, Bryan Drewery wrote: >> On 5/31/16 4:37 PM, Bryan Drewery wrote: >>> On 5/31/16 7:06 AM, David Wolfskill wrote: >>>> Kernel build failed (on laptop). But siince the build machine had succeeded, >>>> I tried "make cleanworld" on the laptop, taht re-tried the build... and >>>> it worked. >>>> --- machine --- >>>> machine -> /root/svn/base/sys/amd64/include >>>> ln: machine/include: File exists >>>> *** [machine] Error code 1 >>>> >>>> make: stopped in /root/svn/base/sys/modules/accf_http >>> >>> >>> I've figured out the root cause for this and will have a fix soon. >>> >> >> Should be fixed by r301088. >> >> If you hit a Linux64.ko issue, you forgot to cleanworld first. Sorry >> about that. I've added in a mitigation to rebuild in this case and am >> working to get a change into bmake to remove the cleanworld need. >> >> The last remaining failure reported is in sys/boot/efi/loader when >> running installworld. >> > > I'm thinking the installworld issue is actually fixed by r301088 as > well. My theory is that the bad machine symlink being created in the > source directory caused the symlink for sys/boot/efi/loader to be > considered modified and it rebuilt on install. > Another reported issue just now is that right after an installworld, everything rebuilds due to changed /bin/sh (-dM flag to make tells you why things rebuild). I'll look into some mitigations for this. -- Regards, Bryan Drewery
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:05 UTC