Bug 244729
| Summary: | Buildworld fails when installed world is built WITHOUT_LEGACY_CONSOLE | ||
|---|---|---|---|
| Product: | Base System | Reporter: | ktullavik |
| Component: | misc | Assignee: | Glen Barber <gjb> |
| Status: | Closed FIXED | ||
| Severity: | Affects Only Me | CC: | 2khramtsov, cperciva, emaste, js, re |
| Priority: | --- | ||
| Version: | CURRENT | ||
| Hardware: | amd64 | ||
| OS: | Any | ||
| See Also: |
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230838 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=254008 |
||
|
Description
ktullavik
2020-03-11 02:26:31 UTC
I just wasted more than a day of my life trying to upgrade from 12.2 to 13.0-RC3 because of this issue. I broke and unbroke my system many times and built world like 10 times before I figured out which src.conf knob was causing this problem. This is going to frustrate the hell out of many users trying to upgrade to 13.0 from source. If WITHOUT_LEGACY_CONSOLE works just fine on 12, there is a reasonable expectation for it to work on 13. I am now trying to reproduce this failure case. In the meantime, could you please provide the contents of your /etc/src.conf and /etc/make.conf? With WITHOUT_LEGACY_CONSOLE=1 in /etc/src.conf, I am unable to reproduce this build failure. (In reply to Glen Barber from comment #3) I presume you need to have a system already installed WITHOUT_LEGACY_CONSOLE to reproduce this - i.e., these do not exist in /usr/sbin kbdcontrol kbdmap moused vidcontrol Confirmed. root@releng1:/usr/src # make -s -j48 buildworld buildkernel ; echo ^G make[1]: "/usr/src/Makefile.inc1" line 339: SYSTEM_COMPILER: Determined that CC=cc matches the source tree. Not bootstrapping a cross-compiler. make[1]: "/usr/src/Makefile.inc1" line 344: SYSTEM_LINKER: Determined that LD=ld matches the source tree. Not bootstrapping a cross-linker. -------------------------------------------------------------- >>> World build started on Tue Mar 23 13:47:59 UTC 2021 -------------------------------------------------------------- -------------------------------------------------------------- >>> Rebuilding the temporary build tree -------------------------------------------------------------- Linking host tools into /usr/obj/usr/src/amd64.amd64/tmp/legacy/bin -------------------------------------------------------------- >>> stage 1.1: legacy release compatibility shims -------------------------------------------------------------- ===> tools/build (obj,includes,all,install) -------------------------------------------------------------- >>> stage 1.2: bootstrap tools -------------------------------------------------------------- --- _bootstrap-tools-link-kbdcontrol --- make[2]: stopped in /usr/src --- _bootstrap-tools --- make[1]: stopped in /usr/src --- buildworld --- make: stopped in /usr/src Committed as 50179c5ec7d09d7b02497caf95dca5d33f93bcd9. Sorry for forgetting to reference the PR. (In reply to Glen Barber from comment #7) Sorry to post somewhat offtopic, but: 'Sponsored by: Rubicon Communications, LLC ("Netgate")' ? Commiting already reported, authored, and reviewed changes by someone and claiming that your employer sponsored this, along without mentioning the people who took their time reporting, investigating where things break, authoring, really? And I no longer have this revision under my profile on Phabricator after you commandeered the revision. I like to keep track of what I've done for FreeBSD. This was most likely an automated preset, but please, pay more attention to this next time. (In reply to Evgeniy Khramtsov from comment #8) I think the "Sponsored by" is correct -- Netgate did sponsor Glen doing the commit. Probably there should have been a "Submitted by" line though. As for the review being commandeered in Phabricator -- that's something which Phabricator does automatically. When FreeBSD started using Phabricator it was only used by committers so that probably made sense at the time; I'll inquire as to whether it's a default which can be changed (especially for reviews opened by non-committers, who will *always* have their work committed by someone else). (In reply to Colin Percival from comment #9) Ping. It's been days since the "MFC after 12 hours" commit, and the one-liner fix, in any way, still isn't in {stable,releng}. (In reply to Evgeniy Khramtsov from comment #10) https://lists.freebsd.org/pipermail/freebsd-current/2021-March/079256.html |
