On 2/4/21 5:54 PM, Pierre Labastie wrote:
On Thu, 2021-02-04 at 17:11 -0500, Scott Andrews wrote:
On 2/4/21 4:46 PM, Pierre Labastie wrote:
On Thu, 2021-02-04 at 21:50 +0100, Frans de Boer wrote:
On 04/02/2021 21:21, Pierre Labastie wrote:
On Thu, 2021-02-04 at 20:03 +0100, Frans de Boer wrote:
On 04/02/2021 16:22, Scott Andrews wrote:
On 2/4/21 9:14 AM, Frans de Boer wrote:
The new instructions for compiling file-5.39 in the first
phase,
are
incomplete. As it turns out, you can only compile file
when
the
configure command is augmented with '--disable-
libseccomp'.
The
file
$LFS/usr/include/libseccomp/seccomp.h does not exist yet,
cause
to
halt compilation with the message that seccomp.h is
missing.
--- Frans.
That should exist as it should have been installed in LFS-
10.0
5.4.1.
BTW I am not a fan of the new "building system" as it takes
excessive
liberties about overwriting things
No, what is installed is /usr/include/seccomp.h, not the one
what
is
needed.
I think this need some clarification on your side: when did you
get
that missing include? When compiling the first pass of file (in
build),
or the second pass?
In the first pass, you depend on what is on the host, it's a
"normal"
compilation on host. It may be that seccomp.h is missing on the
host,
some -dev or -devel package is not installed. I've never seen
that
before.
In the second pass, you depend on what is in $LFS, and yes
seccomp.h is
installed in 5.4.1 (see last line of the instructions):
usr/include
(no /, so relative to the kernel build dir) is copied
recursively
to
$LFS/usr.
Pierre
Pierre,
As stated above, it happens in the first pass.
I have on the host both /usr/include/seccomp.h and
/usr/include/libseccomp/seccomp.h.
However, on the LFS side we only have the first one. And since
the
first
pass is compiled with the LFS includes
It shoudn't, how do you want the build system to know about $LFS if
not
passed --host? The compiler that is used is the one on host: you
can
check in the logs:
checking for gcc... gcc.
The only gcc in the PATH is the one on host (or at least, it should
be
the only one at this stage).
But maybe the problem is that the host /usr/include/seccomp.h tries
to
include /usr/include/linux/seccomp.h, and you do not have that one.
Actually, I do not have /usr/include/libseccomp/seccomp.h on any
host.
On several I only have /usr/include/linux/seccomp.h. On some I have
/usr/include/seccomp.h too.
I think what is needed by file is:
- if the library /usr/lib/libseccomp.so is found on the host and
--disable-libseccomp is not passed:
/usr/include/seccomp.h and /usr/include/linux/seccomp.h (on
the host)
- if the library /usr/lib/libseccomp.so is not found on the host or
--disable-libseccomp is passed:
nothing...
Note that I do not want to point out an error on your side or
whatever:
just try to understand what can go wrong...
Pierre
According to my research.......
/usr/include/seccomp.h is installed by linux-api-headers and only by
linux-api-headers.
This one is installed by libseccomp (in BLFS).
/usr/include/linux/seccomp.h (one more directory in the path) is
installed by linux-api-headers. AFAICT, linux-api-headers installs
nothing directly in /usr/include.
How do I know? My package manager tells me so.
/usr/include/libseccomp/seccomp.h points to a host problem as the
linux headers
Looks like Suse has the seccomp.h header in that directory. So distro
dependent.
install it into /usr/include/ therefore the host system
developers/builders must
have moved to into its own directory.
Do a diff on /usr/include/seccomp.h and
/usr/include/libseccomp/seccomp.h
I believe you will find them one in the same.
As Frans said: problem solved anyway. The file build system is not very
robust: it fails if libraries are installed, but not the corresponding
include files.
Pierre
I missed the "linux" in the path
This is the only seccomp.h on my system
Core/linux_api_headers-20210108-1.armv7hnl.rpm:/usr/include/linux/seccomp.h
--
http://lists.linuxfromscratch.org/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
Do not top post on this list.
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
http://en.wikipedia.org/wiki/Posting_style
