Re: [bitbake-devel] [PATCH] lib/bb/version.py: add from distutils.version

[Tim Orling <ticotimo@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2112 bytes --]

On Fri, Nov 26, 2021 at 8:36 AM Richard Purdie <
richard.purdie@linuxfoundation.org> wrote:

> On Fri, 2021-11-26 at 08:14 -0800, Tim Orling wrote:
> >
> >
> > On Fri, Nov 26, 2021 at 12:21 AM Richard Purdie
> > <richard.purdie@linuxfoundation.org> wrote:
> > > On Thu, 2021-11-25 at 21:18 -0800, Tim Orling wrote:
> > > > In places like sanity.bbclass we currently use
> > > > distutils.version.LooseVersion, but distutils is deprecated in Python
> > > > 3.10 and will be removed in Python 3.12 (~October 2023).
> > > >
> > > > The recommended replacement for for distutils.version.LooseVersion is
> > > > packaging.version.Version, but this implies 'packaging' be installed
> on
> > > > the host. It is also not the same functionality as LooseVersion.
> > > >
> > > > [YOCTO #14610]
> > > >
> > > > Signed-off-by: Tim Orling <timothy.t.orling@intel.com>
> > > > ---
> > > >   lib/bb/version.py | 347
> ++++++++++++++++++++++++++++++++++++++++++++++
> > > >   1 file changed, 347 insertions(+)
> > > >   create mode 100644 lib/bb/version.py
> > >
> > > I'm wondering if we could switch the usage of LooseVersion for
> > > bb.utils.vercmp_string(a, b)?
> > >
> > >
> >
> > I don't know. I was trying to follow any guidance from upstream CPython
> and
> > PEP 632 and this was the simplest way out. In my thread to the
> oe-architecture
> > list Khem thought this was ok, so I just went with it to get the
> distutils
> > deprecation series out into the wild and get eyeballs on it. Much more
> will
> > have to wait until after the YP Summit in terms of my bandwidth.
> >
>
> Ok, I've sent out a patch which replaces the LooseVersion usage with our
> bb.utils function since as far as I can tell it should work fine. It isn't
> quite
> as pretty as the python one but I'm not keen to add new licenses to
> bitbake or
> have multiple version handling functions.
>

Thank you and agreed. I didn’t like adding the new license.

>
> I suspect we could also make that function more pythonic.


I’ll take look later.

>
>
> Cheers,
>
> Richard
>
>

[-- Attachment #2: Type: text/html, Size: 3538 bytes --]