Bug#822504: lintian: false positive spelling-error-in-copyright (duplicate word) involving names and field labels

[Jakub Wilk <jwilk@debian.org>]

Hi Andreas!

* Andreas Beckmann <anbe@debian.org>, 2016-04-25, 03:28:
> this paragraph from debian/copyright triggers a few false positives:
>
> ==========
> Files: debian/source/lintian-overrides
> Copyright: © 2014 Pedro Pérez Pérez
>           © 2015 Wang Wang and Funi
>           © 2016 F. Ree License
> License: GPL-2.0+
> ==========
>
> spelling-error-in-copyright Wang Wang (duplicate word) Wang
> spelling-error-in-copyright License License (duplicate word) License
>
> The copyright field contains names. Names shouldn't be subject to 
> spellchecking.
> And I happen to know a guy from Spain with a similar "duplicate" family 
> name.
>
> Field labels shouldn't be checked either, nor should different fields 
> be concatenated (with their labels inbetween).
>
> (I initially triggered this with
>    Copyright: no copyright notice, see License
>    License: public-domain
>     ...
> )

The spellcheker isn't currently aware of copyright format (or any format 
FWIW), and copyright files aren't required to follow this format anyway. 
This is going to be tough to fix if we want to keep checking for 
duplicated words.

It was me who proposed the duplicated word check (see #800476), but in 
hindsight it might not be such a great idea. :-\ I'd be in favor of 
reverting this change, at least until we find a way to make the check 
more robust.

--
Jakub Wilk