Re: [gentoo-portage-dev] [PATCH 1/2] Split output for repoman checks into file and message

[Brian Dolbec]

On Wed, 19 Feb 2014 10:33:15 -0800
Brian Dolbec <dolsen@g.o> wrote:

> On Wed, 19 Feb 2014 13:10:04 -0500
> Chris Reffett <creffett@g.o> wrote:
>
> > This wraps the output of emerge checks so that a list of length 1-2 is
> > generated. The first element is the file, the second element (optional)
> > is a more descriptive error message. This change will help us eventually
> > introduce more machine-readable output formats.
> > ---
> >  bin/repoman              | 232 +++++++++++++++++++++++------------------------
> >  pym/repoman/utilities.py |  18 +++-
> >  2 files changed, 128 insertions(+), 122 deletions(-)
> >
> > diff --git a/bin/repoman b/bin/repoman
> > index 888892b..3d5dde4 100755
> > --- a/bin/repoman
> > +++ b/bin/repoman
> > @@ -1402,7 +1402,7 @@ for x in effective_scanlist:
> >  		repoman_settings['PORTAGE_QUIET'] = '1'
> >  		if not portage.digestcheck([], repoman_settings, strict=1):
> >  			stats["manifest.bad"] += 1
> > -			fails["manifest.bad"].append(os.path.join(x, 'Manifest'))
> > +			fails["manifest.bad"].append([os.path.join(x, 'Manifest')])
> >  		repoman_settings.pop('PORTAGE_QUIET', None)
> >
>
> This looks so,so to me.  I think you are much better off using a
> namedtuple class for these errors instead.  They have built-in
> formatted printing, etc.
>
> from collections import namedtuple
>
> and you can have pre-defined named tuple classes that have the error
> message already embedded.
>
> for some examples see the pyGPG ones I created for gpg data mining:
>
> https://github.com/dol-sen/pyGPG/blob/master/pyGPG/legend.py
>
> NOTE: CLASSES is a list of tuples that have the info to define the
> classes that subclass the namedtuple class.  They are initialized by
> the code at the bottom of the page when the page is first loaded into
> memory.  This format saves writing out each individual class by hand
> and makes changes easy.  It also reduced the size of the file to about
> 1/3.  I have done similar to tis in 3 modules in the catalys rewrite
> as well.
>
> The data is then available in many ways and will have a consistent
> output.

for smaller simpler examples:

http://git.overlays.gentoo.org/gitweb/?p=proj/catalyst.git;a=blob;f=catalyst/hash_utils.py;h=cd31ad3edbdd412c0da4d968596777a71fbd4beb;hb=refs/heads/3.0

http://git.overlays.gentoo.org/gitweb/?p=proj/catalyst.git;a=blob;f=catalyst/contents.py;h=1a9ed28a58cc63ed954ca8bdbcd1b594e8a7f2e5;hb=refs/heads/3.0


-- 
Brian Dolbec <dolsen>