[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: icc-profiles_2.2_source.changes REJECTED



Le lundi 27 février 2023, 12:11:27 UTC Jonas Smedegaard a écrit :
Hi jonas,

I have just checked the source code of lintian. Could you double check your package and create a simple test case ?

According to:
https://salsa.debian.org/lintian/lintian/-/blob/master/lib/Lintian/Check/Files/NonFree.pm#L91
The test should not raise

Bastien
> Quoting Simon McVittie (2023-02-27 13:04:25)
> > On Mon, 27 Feb 2023 at 12:22:34 +0100, Jonas Smedegaard wrote:
> > > I am not convinced, however, that this issue is a bug in lintian:
> > > The testcase you ask for is the actual files in the icc-profiles source
> > > package which already is already correctly identified by lintian.  I.e.
> > > issue is not that some different files get misdetected but instead that
> > > lintian _correctly_ identifies the files in this non-free package and
> > > _correctly_ classifies those as unsuitable in main
> > 
> > That's correct to a point, but it seems wrong for lintian to be emitting a
> > tag that means "this package is unsuitable for main and that's a problem"
> > for a package in non-free. Yes, we know it's unsuitable for main, and
> > the maintainer has acknowledged that by uploading it to non-free... it
> > doesn't seem like there is any benefit to having an unfixable Lintian
> > error as well.
> > 
> > Or if the interpretation of the tag is "you should use the copy of this
> > file from icc-profiles instead of shipping your own copy", then the lintian
> > check should have a special case to silence that tag when the package being
> > checked *is* icc-profiles.
> > 
> > There are similar special cases in other parts of Lintian. For example,
> > it looks for embedded code copies of zlib, but that check has a special
> > case to avoid it being triggered by zlib itself, because obviously zlib
> > should be allowed to ship zlib code; and similarly checks for a bundled
> > font like Deja Vu shouldn't trigger for fonts-dejavu itself.
> 
> Ah, I was unaware that lintian at other places includes exceptions like
> that, an that lintian is expected to do reliable assessments in this
> kind of tests.
> 
> When that's the case, then please, Bastien (or others skilled with
> internals of lintian) please refine the tests for icc-profiles contents
> to exclude icc-profiles package itself.
> 
> (I'll file a bugreport against lintian if I notice no actions from this)
> 
> Kind regards,
> 
>  - Jonas
> 
> 

Attachment: signature.asc
Description: This is a digitally signed message part.


Reply to: