On 2/3/2020 5:17 PM, Ken Moffat via blfs-support wrote:
On Mon, Feb 03, 2020 at 02:29:54PM -0700, Alan Feuerbacher via blfs-support
wrote:
After I began installing BLFS packages (in systemd development), I
found that executing pip, pip2 or pip3 gave a long error message
something like this:
(snipping to the part that jumped out - for me if not for you - when
I read it)
subprocess.CalledProcessError: Command '('lsb_release', '-a')' returned
non-zero exit status 1.
I'm snipping the traceback because I generally find python traces to
be very hard to understand, although I'm sure that some pythonistas
like them.
I will hazard a guess that either you don't have an lsb-release file
(see BLFS chapter 12, LSB-Tools-0.5 - the configuration was done in
LFS chapter 9 section 9.1.
Right. Somehow, in my 20-hour marathon rebuild session, I missed that
section altogether. :-(
So I tried to rebuild Python 3 and Python 2 according to BLFS book
instructions. But when I executed "make install" I got essentially the
same error message as above. After looking on the Net for some
guidance, and trying a variety of things to figure out what the error
is, I tried going back to the LFS tool chain with the following, as
nonprivileged user alan, using ideas I recently learned from Bruce:
Sometimes, rebuilding is necessary. Other times it just delays
finding the problem. I think this is an example of the latter,
Your second failure again mentioned lsb_release.
Right again. See below.
At worst I could go back and start from scratch with LFS yet again on
another fresh hard drive, but I really don't want to do that, since it
requires about 30 hours to get to the point of building Python-2.7.17.
I'd appreciate any help on this.
Alan
Quite why pip wants to look at lsb-release is beyond me, but I
suspect it looks for known distros and perhaps applies 'quirks' to
some of them.
What do you have in /etc/lsb-release ?
See my reply to Don Cross a few minutes ago.
As I told Don, I think I've found the actual problem: Python-3.8.1 fails
to install two binaries that the LFS book says are created:
/usr/bin/pip3.8 is not created at all, but a symlink /usr/bin/pip3 ->
pip3.8 is created by a manual step. That's why trying to execute it
fails, since the symlink points to a nonexistent file.
/usr/bin/pyvenv and /usr/bin/pyvenv-3.8 are not created at all.
I have no idea why those binaries are missing. I deleted all the
python3-related binaries from /usr/bin and rebuilt python-3.8.1, then
looked at what new binaries showed up. All but the above-mentioned ones
were rebuilt.
Whew! I was about to look for my samurai sword to commit seppuku at the
thought of starting from scratch again.
Alan
--
http://lists.linuxfromscratch.org/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
