Bug#798900: lintian: false positive: source-is-missing for non-minified JS files
Am 20.09.2015 um 17:44 schrieb Bastien ROUCARIES:
> On Thu, Sep 17, 2015 at 9:36 AM, Ole Streicher <olebole@debian.org> wrote:
>> control: affects -1 python-astropy
>>
>> Same for astropy; also probably because of a long line:
>>
>> http://lintian.debian.org/maintainer/debian-astro-maintainers@lists.alioth.debian.org.html#python-astropy
>>
>> Questioned source file:
>>
>> https://sources.debian.net/src/python-astropy/1.0.4-1/astropy/extern/js/jquery.dataTables.js
>>
>> The heuristics is obviously wrong.
>>
>> IMO, if there is a good chance for false positives, this
>> tag should not be marked as "Error", but as "Experimental".
> patch are welcome.
I am not familar with the lintian sources, but I guess that changing the
tag should be quite simple and not require a patch.
> In your case they are a line > 1024.... Likely the
> /*globals $,
>
> Are you sure this line was generated by hand ? I think this line was
> generated by a script and thus this file is not really the prefered
> source of modification (it is bordeline).
Could you please specify a use case in which you would prefer another
source?
An automatically inserted comment line does not necessarily make a file
"generated", for example:
* Sources that are handled in CVS or SVN may contain automatic content
(cvsid and similar)
* debian/changelog has automatically inserted content: the author and
the date if one uses "dch", and some workflows even auto-update
debian/changelog from git commit messages.
* debian/patches/* is completely auto-generated, and obviously *not*
someting one would prefer to edit by hand. Sometimes it even contains
messages from git (if they were not created with quilt, but with "git
diff"); for these messages the source is usually not included at all.
Would you f.e. also try to invalidate /usr/include/linux/capi.h as being
non-source since it contains the following generated line?
/* $Id: capi.h,v 1.4.6.1 2001/09/23 22:25:05 kai Exp $
What is the difference to the one liner in the questioned file?
Sorry, I even don't see that having a single (even long) comment line
auto-generated in a file is necessarily a "borderline". Even if you
would estimate it as "borderline" (which would consequently as well
apply to the CVS Ids), this would not qualify for the "error" tag, but
for the "pedantic" (since it means that people may disagree). I would,
however, propose "experimental", since the heuristics is just not mature
yet.
Best regards
Ole
Reply to: