from
Sylvain Thénault <sylvain.thenault at logilab dot fr>
to
Maarten ter Huurne <maarten.ter.huurne at philips dot com>
cc
Sylvain Thénault <sylvain.thenault at logilab dot fr>
subject
Re: [Python-projects] False positive in static analysis
ofclassmembers.
date
On Thursday 09 March à 18:26, Maarten ter Huurne wrote:2006/03/09 18:32
> syt@logilab.fr wrote on 2006-03-09 03:16:10 PM:
>
> > On Thursday 09 March à 15:10, Maarten ter Huurne wrote:
> > > Indeed it does work if I put "--disable-msg" first.
> > >
> > > However, there is another problem: when running pylint, disabled
> messages
> > > from the command line and from the rcfile are merged (so if a message
> is
> > > disabled in either location, it is disabled for the run).
> >
> > yes. If you want to enable a message disabled in the rc file, you have
> > to use the --enable-msg option to override this pb.
>
> Indeed, this works exactly as I would expect. The unexpected thing is the
> one described below.
>
> > > But when using
> > > "--generate-rcfile", the messages that were already disabled in the
> rcfile
> > > do not show up in the output, only the messages disabled from the
> command
> > > line are in the output.
> >
> > hum, this is strange indeed. I've added this to our tracker to take a
> > look on it latter.
>
> I'm not sure if this is related, but the rcfile documentation for the
> "ignore" options says it can be specified multiple times:
well, this is true for command line option, no for the configuration
file.
> # Add <file or directory> to the black list. It should be a base name, not
> a
> # path. You may set this option multiple times.
>
> However, if you specify it like this:
>
> ignore=CVS
> ignore=source.py
>
> ...only one of them takes effect (I forgot if it was the first or the
> last). To specify multiple ignores, they should be comma separated:
>
> ignore=CVS,source.py
>
> So it seems the pylint config data structure only accepts one instance of
> each key and others are lost. Maybe the same thing happens with multiple
> "disable-msg" definitions?
hum, I don't think so, but I'll check...
--
Sylvain Thénault LOGILAB, Paris (France).
http://www.logilab.com http://www.logilab.fr http://www.logilab.org
