Re: [PATCH 2/4] gnu: Add r-bigmemory-sri.

[Roel Janssen]

Ricardo Wurmus writes:

> Roel Janssen <address@hidden> writes:
>
>> Ricardo Wurmus writes:
>>
>>> Roel Janssen <address@hidden> writes:
>>>
>>>> Roel Janssen writes:
>>>>
>>>>> Ricardo Wurmus writes:
>>>>>
>>>>>> Roel Janssen <address@hidden> writes:
>>>>>>> +    (description "This package provides a shared resource interface 
>>>>>>> for the
>>>>>>> +bigmemory and synchronicity packages.")
>>>>>>> +    (license (list license:lgpl3 license:asl2.0))))
>>>>>>
>>>>>> What does this list mean?
>>>>>> Also: is this LGPL3+ or LGPL3 only?
>>>>>
>>>>> The CRAN page lists LGPL3 explicitly, but that could be imprecise ...
>>>>> The source code package does not contain any other license indication
>>>>> than waht is stated in the DESCRIPTION file (which states LGPL3 and
>>>>> Apache Software License 2.0).
>>>>>
>>>>> See:
>>>>>   https://cran.r-project.org/web/packages/bigmemory.sri/
>>>>>
>>>>> So, I think the only thing I can do is just follow what has been stated,
>>>>> which is LGPL3 (precisely this) and Apache Software License 2.0.
>>>>
>>>> I don't know how to proceed now.  I think it's fine as the list of
>>>> licenses is the list of licenses they provide.
>>>>
>>>> Are these licenses incompatible?  If so, then there's nothing I can do
>>>> either, because these are the licenses that are provided..
>>>
>>> Usually, what we do for R is to assume “or later” because that’s how
>>> things are usually done on CRAN.  (They also automatically expand
>>> license declarations.)
>>
>> Well I don't think we can do that in this case because that's not what
>> the license field says.  In the code there's no license at all, so that
>> makes it even more difficult.
>
> You’re right.  I misremembered.  It’s only these joint license
> declarations like “GPL-2 | GPL-3” that effectively mean “or later”.  (It
> is impossible to express “or later” in canonical R license fields.)
>
>> I guess this is about the possible license incompatibility between LGPLv3
>> and Apache?  I tried to explain that in any case, there's nothing I can
>> do about it anyway..
>
> Actually, the declaration in this package means “either this or that”
> license.
>
> See https://cran.r-project.org/doc/manuals/R-exts.html#The-DESCRIPTION-file
>
>     “The mandatory ‘License’ field in the DESCRIPTION file should
>      specify the license of the package in a standardized
>      form. Alternatives are indicated via vertical bars.”
>
> Your patch is fine if you add a comment above the license field that
> states that these one of these two licenses may be choosen.

Right.  Would the following patch be alright then:

modified   gnu/packages/statistics.scm
@@ -2787,6 +2787,26 @@ Fourier transform, fuzzy clustering, support vector 
machines, shortest path
 computation, bagged clustering, naive Bayes classifier, and more.")
     (license license:gpl2+)))
 
+(define-public r-bigmemory-sri
+  (package
+    (name "r-bigmemory-sri")
+    (version "0.1.3")
+    (source
+     (origin
+       (method url-fetch)
+       (uri (cran-uri "bigmemory.sri" version))
+       (sha256
+        (base32 "0mg14ilwdkd64q2ri9jdwnk7mp55dqim7xfifrs65sdsv1934h2m"))))
+    (properties
+     `((upstream-name . "bigmemory.sri")))
+    (build-system r-build-system)
+    (home-page "http://cran.r-project.org/web/packages/bigmemory.sri";)
+    (synopsis "Shared resource interface for the bigmemory package")
+    (description "This package provides a shared resource interface for the
+bigmemory and synchronicity packages.")
+    ;; Users can choose either LGPLv3 or ASL2.0.
+    (license (list license:lgpl3 license:asl2.0))))
+
 (define-public r-nmf
   (package
     (name "r-nmf")

Kind regards,
Roel Janssen