[OE-core] [PATCH] oeqa.utils.metadata: cope with invalid lines in os-release file

[Markus Lehtonen]

On 26/04/2017, 17.47, "Joshua Lock" <joshua.g.lock at linux.intel.com> wrote:

    On Wed, 2017-04-26 at 17:39 +0300, Markus Lehtonen wrote:
    > Don't crash if every line in /etc/os-release does not adhere to the
    > expected "key=val" format. E.g. CentOS 7 has empty lines in the file.
    > 
    > Signed-off-by: Markus Lehtonen <markus.lehtonen at linux.intel.com>
    > ---
    >  meta/lib/oeqa/utils/metadata.py | 6 ++++--
    >  1 file changed, 4 insertions(+), 2 deletions(-)
    > 
    > diff --git a/meta/lib/oeqa/utils/metadata.py
    > b/meta/lib/oeqa/utils/metadata.py
    > index cb81155e54..d291ddb960 100644
    > --- a/meta/lib/oeqa/utils/metadata.py
    > +++ b/meta/lib/oeqa/utils/metadata.py
    > @@ -20,8 +20,10 @@ def get_os_release():
    >          return None
    >      with open(os_release_file) as fobj:
    >          for line in fobj:
    > -            key, value = line.split('=', 1)
    > -            data[key.strip().lower()] = value.strip().strip('"')
    > +            split = line.split('=', 1)
    > +            if len(split) == 2:
    > +                key, value = split
    > +                data[key.strip().lower()] = value.strip().strip('"')
    
    We have a function to read os-release in oe.lsb, return_dict_osr().
    
    It handles this situation slightly differently, but the real reason I'm
    pointing this out is that it seems like we could factor out a shared
    function to be used both here and in oe.lsb ?


Thanks for spotting this. Makes sense to me. I'll send a new patch merging the two functions shortly.

Thanks,
   Markus