[haiku-development] Haiku build fails on HAIKU_NO_DOWNLOADS: updated
- From: "Thomas Mueller" <mueller6723@xxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Mon, 18 Aug 2014 05:54:58 +0000
from Ingo Weinhold and my last message:
> On 17.08.2014 04:48, Thomas Mueller wrote:
> > There is another problem, HOST_SFDISK ?= sfdisk ; which does not
> > exist in NetBSD or FreeBSD base system.
> I think it is (was?) used in the ARM build. It isn't used in the regular
> x86* builds, so you can ignore it, if you want to focus on those.
Seeing HOST_SFDISK in BuildConfig made me nervous.
I don't have any ARM devices currently to test on.
Platforms of interest are i386 and amd64 (x86 and x86_64).
> > NetBSD has /sbin/fdisk in base, but no *fdisk* in pkgsrc.
> > FreeBSD ports include sysutils/linuxfdisk , then I would define
> > HOST_SFDISK = /usr/local/bin/sfdisk-linux ;
> > Question is whether I need fdisk or sfdisk to build Haiku bootable
> > image (anyboot or whatever else).
> > Also, does jam have an equivalent of "make clean" ?
> There's a "clean" target as well, but it isn't implemented to clean
> everything. I usually just delete manually from the generated directory
> what I want to rebuild. Removing everything but "build",
> "cross-tools-*", and "Jamfile" will reconstruct the post-configure
> state. I normally also keep "download" to avoid re-downloading the
> pre-built packages.
Manual deletion was my approach too, but of course not deleting successfully
built cross-tools-*, Jamfile, and build.
> > /home/nbarlene/haiku/haiku/src/build/libroot/function_remapper.cpp: In
> > function 'int renameat(int, const char*, int, const char*)':
> > /home/nbarlene/haiku/haiku/src/build/libroot/function_remapper.cpp:272:1:
> > warning: 'int renameat(int, const char*, int, const char*)': visibility
> > attribute ignored because it [-Wattributes]
> > renameat(int fromFD, const char* from, int toFD, const char* to)
> ^
> Something seems to be off with the warning message (copy and paste
> issue?). Since this code isn't built with -Werror, there must be an
> actual error as well or the command wouldn't fail.
No copy and paste, I used, in vi editor,
:r !tail -n 13 /home/nbarlene/haiku/haiku/nbgenerated/anyboot.log6
But I must bear in mind that, since NetBSD-current amd64 (6.99.44) became
corrupted, I had to build NetBSD-current anew, this time it was 7.99.1.
I rsync'ed gcc-aux 4.7.3, but that was likely corrupted in part, and common
sense suggests it would be foolhardy to continue using it. I think I'll delete
the copy.
I succeeded in my aim to test with GNU checksum tools such as I would get with
coreutils on Linux.
Tom
Other related posts:
