Re: Apparent build race(s), r315238 -> r315298

From: Ngie Cooper (yaneurabeya) <"Ngie>
Date: Wed, 15 Mar 2017 11:21:40 -0700
> On Mar 15, 2017, at 11:09, Bryan Drewery <bdrewery_at_FreeBSD.org> wrote:
> 
> On 3/15/2017 10:04 AM, David Wolfskill wrote:
> 
>> --------------------------------------------------------------
>>>>> stage 1.1: legacy release compatibility shims
>> --------------------------------------------------------------
>> cd /usr/src; MAKEOBJDIRPREFIX=/usr/obj/usr/src/tmp
> 
> Ok it is trying to use MAKEOBJDIRPREFIX=/usr/obj
> 
>> Building /common/S4/obj/usr/src/gnu/usr.bin/gdb/gdbtui/gdbtui
> 
> But all of the 'Building' lines are tryign to use
> MAKEOBJDIRPREFIX=/common/S4/obj/ ...
> 
>> --- all_subdir_cddl ---
>> --- all_subdir_cddl/usr.bin ---
>> --- all_subdir_cddl/usr.bin/zinject ---
>> ===> cddl/usr.bin/zinject (all)
>> --- all_subdir_gnu ---
>> --- gdbtui ---
>> cc: error: no such file or directory: '/usr/obj/usr/src/gnu/usr.bin/binutils/libbfd/libbfd.a'
> 
> But then complains about a missing file in MAKEOBJDIRPREFIX=/usr/obj
> 
>> *** [gdbtui] Error code 1
>> 
>> bmake[6]: stopped in /usr/src/gnu/usr.bin/gdb/gdbtui
>> .ERROR_TARGET='gdbtui'
>> .ERROR_META_FILE='/common/S4/obj/usr/src/gnu/usr.bin/gdb/gdbtui/gdbtui.meta'
>> .MAKE.LEVEL='6'
>> MAKEFILE=''
>> .MAKE.MODE='meta missing-filemon=yes missing-meta=yes silent=yes verbose'
>> .CURDIR='/usr/src/gnu/usr.bin/gdb/gdbtui'
>> .MAKE='/usr/obj/usr/src/make.amd64/bmake'
> ...
>> .OBJDIR='/usr/obj/usr/src/gnu/usr.bin/gdb/gdbtui'
> 
> And says here that there is an existing directory for
> MAKEOBJDIRPREFIX=/usr/obj for /usr/src/gnu/usr.bin/gdb/gdbtui, which is
> throwing off its paths for libbfd.a.
> 
> ...
>> MAKEOBJDIRPREFIX='/usr/obj'
> 
> And the meta error agrees it wants MAKEOBJDIRPREFIX=/usr/obj
> 
> 
> So where is /common/S4/obj coming from?
> 
> Is there a symlink involved here for /usr/obj?

	^/head_at_r315175 doesn’t look like it’s causing a problem here. I don’t think my commit here recently (^/head_at_r313650) was a problem, but I wonder if it exposed a race or incomplete dependency logic…
Thanks!
-Ngie

Received on Wed Mar 15 2017 - 17:21:42 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:10 UTC