New Build System Status
Chris Johns
chrisj at rtems.org
Tue Nov 26 01:11:29 UTC 2019
On 26/11/19 11:55 am, Joel Sherrill wrote:
> On Mon, Nov 25, 2019, 6:30 PM Chris Johns <chrisj at rtems.org
> <mailto:chrisj at rtems.org>> wrote:
> On 26/11/19 10:46 am, Joel Sherrill wrote:
> > On Mon, Nov 25, 2019 at 5:28 PM Chris Johns <chrisj at rtems.org
> <mailto:chrisj at rtems.org>
> > <mailto:chrisj at rtems.org <mailto:chrisj at rtems.org>>> wrote:
> > On 26/11/19 9:57 am, Joel Sherrill wrote:
> > > On Mon, Nov 25, 2019 at 3:19 PM Chris Johns <chrisj at rtems.org
> <mailto:chrisj at rtems.org>
> > <mailto:chrisj at rtems.org <mailto:chrisj at rtems.org>>
> > > <mailto:chrisj at rtems.org <mailto:chrisj at rtems.org>
> <mailto:chrisj at rtems.org <mailto:chrisj at rtems.org>>>> wrote:
> > >
> > > On 26/11/19 8:05 am, Joel Sherrill wrote:
> > > > On Mon, Nov 25, 2019 at 2:12 AM Sebastian Huber
> > > > <sebastian.huber at embedded-brains.de
> <mailto:sebastian.huber at embedded-brains.de>
> > <mailto:sebastian.huber at embedded-brains.de
> <mailto:sebastian.huber at embedded-brains.de>>
> > > <mailto:sebastian.huber at embedded-brains.de
> <mailto:sebastian.huber at embedded-brains.de>
> > <mailto:sebastian.huber at embedded-brains.de
> <mailto:sebastian.huber at embedded-brains.de>>>
> > > <mailto:sebastian.huber at embedded-brains.de
> <mailto:sebastian.huber at embedded-brains.de>
> > <mailto:sebastian.huber at embedded-brains.de
> <mailto:sebastian.huber at embedded-brains.de>>
> > > <mailto:sebastian.huber at embedded-brains.de
> <mailto:sebastian.huber at embedded-brains.de>
> > <mailto:sebastian.huber at embedded-brains.de
> <mailto:sebastian.huber at embedded-brains.de>>>>>
> > > > wrote:
> > > >
> > > > On 25/11/2019 08:55, Chris Johns wrote:
> > > > >
> > > > > On 25/11/19 5:28 pm, Sebastian Huber wrote:
> > > > >>
> > > > >> On 24/11/2019 23:25, Chris Johns wrote:
> > > > >>>>> Should there be a note or something about waf needing
> > python and we
> > > > recommend
> > > > >>>>> python3? Plus waf needs a `python` installed and not
> just
> > > `python2` or
> > > > >>>>> `python3`?
> > > > >>>> I think this belongs to the Host Computer section.
> The quick
> > > start uses the
> > > > >>>> RSB, so if you managed to build the tools, you must
> have a
> > > working Python.
> > > > >>>> The RSB uses Python and the RTEMS Tools use waf.
> > > > >>> The RSB can use python2 or python3 without a python. What
> > about a note
> > > > to say
> > > > >>> ... "Waf uses python and you need to make this command
> available
> > > on your
> > > > system".
> > > > >> Now I am a bit confused. I thought every UNIX-like system
> > provides a
> > > > >>
> > > > >> #!/usr/bin/env python
> > > > >>
> > > > >> which maps to Python 2 or 3?
> > > > >>
> > > > > Apparently on some distros the python executable is not
> > automatically
> > > > installed
> > > > > with a suitable python2 or python3. I am not aware of the
> > specifics of
> > > > this but
> > > > > it was raised as an issue that I fixed in the RSB:
> > > > >
> > > > > https://devel.rtems.org/ticket/3537
> > > >
> > > > I have never seen a system in which
> > > >
> > > > #!/usr/bin/env python2
> > > >
> > > > or
> > > >
> > > > #!/usr/bin/env python3
> > > >
> > > > works, but not
> > > >
> > > > #!/usr/bin/env python
> > >
> > > I agree however this is the post from Joel about a class he held ...
> > >
> > > https://lists.rtems.org/pipermail/devel/2018-October/023150.html
> > >
> > > and I posted the change here ...
> > >
> > > https://lists.rtems.org/pipermail/devel/2018-October/023296.html
> > >
> > > > How many Python scripts would stop to work in this case?
> > However, the
> > > > availability of a python command is not mandatory
> according to:
> > > >
> > > > https://www.python.org/dev/peps/pep-0394/
> > >
> > > I am aware of this and I support it however Linux distros seems to
> > feel there
> > > are cases for exceptions, we have this one and there is also pax.
> > >
> > > The pax and this python case are similar, do we say "We are correct,
> > fix your
> > > host or raise a bug on your distro" or do we attempt to smooth over
> > the cracks
> > > and attempt to just work?
> > >
> > > In the case of ubuntu, maybe someone with some experience with
> it can
> > see what
> > > is needed and update the documentation?
> > >
> > > > I think we should note in the "Host Computer" section that a
> > "python"
> > > > command is mandatory for RTEMS and remove the RSB extra stuff.
> > >
> > > The argument was presented to me that the RSB should just work
> even in
> > this case
> > > and so that is what I did. I am fine to remove that support, it
> is an
> > > average hack.
> > >
> > > > Chris can answer to that.
> > >
> > > Done.
> > >
> > > > But on Centos 7, we have:
> > > >
> > > > $ type python
> > > > python is /usr/bin/python
> > > > $ python --version
> > > > Python 2.7.5
> > > > $ type python2
> > > > python2 is /usr/bin/python2
> > > > $ type python3
> > > > bash: type: python3: not found
> > > > $ env python
> > > > Python 2.7.5 (default, Aug 7 2019, 00:51:29)
> > > > [GCC 4.8.5 20150623 (Red Hat 4.8.5-39)] on linux2
> > > > Type "help", "copyright", "credits" or "license" for more
> information.
> > > >>>>
> > > >
> > > > and if I activate the SCL for Python 3:
> > > >
> > > > $ scl enable rh-python36 bash
>
> >
> > >
> > > >
>
> >
> > > > [joel at localhost rtems-work]$ type python
>
> >
> > >
> > > >
>
> >
> > > > python is /opt/rh/rh-python36/root/usr/bin/python
>
> >
> > >
> > > >
>
> >
> > > > [joel at localhost rtems-work]$ python --version
>
> >
> > >
> > > >
>
> >
> > > > Python 3.6.3
> > >
> > > And on ubuntu? The original post was about that distro.
> > >
> > >
> > > Welcome to Ubuntu 18.04.3 LTS (GNU/Linux 4.15.0-66-generic x86_64)
>
> >
> > >
> > > ....
> > > $ type python
> > > python is /usr/bin/python
> > > $ python --version
> > > Python 2.7.15+
> > > $ env python
> > > Python 2.7.15+ (default, Oct 7 2019, 17:39:04)
> > > [GCC 7.4.0] on linux2
> > > Type "help", "copyright", "credits" or "license" for more information.
> > >>>>
> > > $ python2
> > > Python 2.7.15+ (default, Oct 7 2019, 17:39:04)
> > > [GCC 7.4.0] on linux2
> > > Type "help", "copyright", "credits" or "license" for more information.
> > >>>>
> > > $ python3
> > > Python 3.6.8 (default, Oct 7 2019, 12:59:55)
> > > [GCC 8.3.0] on linux
> > > Type "help", "copyright", "credits" or "license" for more information.
> > >>>>
> >
> > I am sorry but I am not sure how this helps the discussion? Your
> original post
> > last year raised the issue of no python being installed on ubuntu and
> that lead
> > to the change in the RSB.
> >
> > I wish I had included the Ubuntu version in the original post. msys2 and
> cygwin
> > always seem to reflect the day you install them to me. The best I could do is
> > report on the msys2 on my laptop and the Ubuntu we use for testing appear
> to be OK.
> >
> > I have no way to reproduce what happened in the class I sent an email about.
> > I assure you that it happened though.
>
> I am sure it did happen. I reported at the time the python symlink is part of
> the python-minimal package. I have no idea how this sits in that distro's
> packaging structure and how it can happen.
>
> What does you version show for ...
>
> apt-file search /usr/bin/python | grep "/usr/bin/python$"
>
> ... ?
>
> > On the master, I suppose assuming "python" exists is OK as an experiment
> as long
> > as it is assumed it may be Python2 or Python3. We should be prepared for
> someone
> > to have the situation where it doesn't work out though and need to revisit
> this.
>
> Are you saying it is OK to remove the support I added from the RSB and
> rtems-tools?
>
>
> I am saying I have no system to prove it right now. I think it is still needed
> but can't reproduce it.
>
> I don't like removing it quietly without a replacement to detect and report
> something to a user that X occurred, please report so we know what hosts this
> happens on.
What happens with the new build system and waf as it now requires a valid
python? This is the point Sebastian is making and it is a valid one. We also
have rtems-tools and libbsd which already exist and have the same issue.
Chris
More information about the devel
mailing list