Re: Another recent WITH_META_MODE= gotcha ": handling svn commit: r334008 - head/bin/sh

2018-05-25 Thread Bryan Drewery
On 5/25/2018 2:49 AM, M - Krasznai András wrote:
> Hi
> 
> I found another problem with compiling FreeBSD-CURRENT
> 
> when I tried to compile GENERIC instead of my custom I run into the following:
> 
> make installworld fails with 
> 
> /bin/sh: /usr/obj/usr/src/amd64.amd64/rescue/rescue/rescue: not found
> rescue/sh check failed, installation aborted. 
> 
> I used originally a customized src.conf which - among others - specifies 
> WITHOUT_RESCUE.
> When trying to compile GENERIC I removed src.conf and I thought I would get a 
> complete FreeBSD-CURRENT with GENERIC, with rescue files, etc. 
> 

Unclear to me. You did buildworld with WITHOUT_RESCUE=yes set and then
removed 'WITHOUT_RESCUE' from your src.conf (by removing src.conf), and
then did buildworld and then installworld and hit this issue? All with
WITH_META_MODE=yes? Or did you disable META_MODE at some point?

> The source tree is at r334200
> 
> 
> best regards
> 
> András Krasznai


-- 
Regards,
Bryan Drewery



signature.asc
Description: OpenPGP digital signature


You might want to be careful around r334147...

2018-05-25 Thread David Wolfskill
I had no issues on my build machine (head/amd64; GENERIC kernel)
performing src-based upgrades from r333974 -> r334147 -> r334205
(first step yesterday; second today).

On my laptop (also head/amd64; lightly-customize kernel that includes
GENERIC), I was able to go from r333974 -> r334147, booted r334147
OK and did a quick smoke-test (login via xdm; stuff seemed to work).

Today, however, after updating source to r334205, the laptop locked up
hard (only way out I could find was power-cycle) twice, and performed a
spontaneous reboot -- each during "make buildworld."

As things presently stand, I have reverted to the r333974 kernel
(built the day before yesterday) and have started the build for
r334205 over with an empty /usr/obj.

Peace,
david
-- 
David H. Wolfskill  da...@catwhisker.org
Trump's EPA excludes reporters from a meeting on contaminants -- why?

See http://www.catwhisker.org/~david/publickey.gpg for my public key.


signature.asc
Description: PGP signature


RE: Another recent WITH_META_MODE= gotcha ": handling svn commit: r334008 - head/bin/sh

2018-05-25 Thread M - Krasznai András
Hi

I found another problem with compiling FreeBSD-CURRENT

when I tried to compile GENERIC instead of my custom I run into the following:

make installworld fails with 

/bin/sh: /usr/obj/usr/src/amd64.amd64/rescue/rescue/rescue: not found
rescue/sh check failed, installation aborted. 

I used originally a customized src.conf which - among others - specifies 
WITHOUT_RESCUE.
When trying to compile GENERIC I removed src.conf and I thought I would get a 
complete FreeBSD-CURRENT with GENERIC, with rescue files, etc. 

The source tree is at r334200


best regards

András Krasznai


-Eredeti üzenet-
Feladó: owner-freebsd-curr...@freebsd.org 
[mailto:owner-freebsd-curr...@freebsd.org] Meghatalmazó Bryan Drewery
Küldve: 2018. május 24. 20:50
Címzett: Mark Millard; FreeBSD Current
Másolatot kap: cy.schub...@cschubert.com; O. Hartmann
Tárgy: Re: Another recent WITH_META_MODE= gotcha ": handling svn commit: 
r334008 - head/bin/sh

On 5/21/2018 10:17 PM, Mark Millard wrote:
> Attempting to build head -r334014 from -r333947 includes the code from 
> -r334008, and for which I get:
> 
> --- parser.o ---
> /usr/src/bin/sh/parser.c:1440:9: error: use of undeclared identifier 'CQNL'
> case CQNL:
>  ^
> 1 error generated.
> 
> for what was added in -r334008.
> 
> # grep -r CQNL /usr/src/* | more
> /usr/src/bin/sh/mksyntax.c: { "CQNL",   "newline character in quotes" 
> },
> /usr/src/bin/sh/mksyntax.c: add("\n", "CQNL");
> /usr/src/bin/sh/mksyntax.c: add("\n", "CQNL");
> /usr/src/bin/sh/mksyntax.c: add("\n", "CQNL");
> /usr/src/bin/sh/parser.c:   case CQNL:
> /usr/src/contrib/libarchive/libarchive/test/test_read_format_rar_binar
> y_data.rar.uu:M+M7)?9;.A%CQNL-+4::&_UJQ*P"PPW:>)I5*8)7^>6(\]!Q"^9YCE>>
> `9OG8
> 
> Apparently this is something WITH_META_MODE= does not deal with.
> Rebuilding after:
> 
> rm -fr /usr/obj/amd64_clang/*
> 
> (here I normally have the build materials) built fine.
> 
> (The "Making top.local.h from /usr/src/contrib/top/top.local.hs"
> change to no longer have top.local.h or top.local.hs was another
> example: for a while top.local.h was still referenced and old ones 
> would still be used if present.)
> 

It was a bug with the rescue build with dependency handling. I've fixed it in 
r334177.


--
Regards,
Bryan Drewery

___
freebsd-current@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"


Re: Boot USB memstick with MBR (WAS: Re: [RFC] Deprecation and removal of the drm2 driver)

2018-05-25 Thread Philip Homburg
>Can you download the image from
>https://people.freebsd.org/~emaste/mini-image.amd64.xz, uncompress and
>write it to a USB stick and try on this system? This is a
>MBR-partitioned dual-mode test image. It's not an installer - it
>should just boot to a login prompt - but can be used to test this
>scheme.

I have bad news and some good news.

The bad news is that with this image the USB stick doesn't get recognized as
a boot device. Same as with the 11.2-BETA2 images.

The good news is that if I completely recreate the MBR, keeping the layout of
the partitions the same, then the USB stick is recognized.

The weird news is that it then proceeds to boot from harddisk. But that may
be an issue with the boot code I put on.

So the issue is limited to the MBR. 

I'll try to figure out what exactly makes the difference. 


___
freebsd-current@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"